diff --git a/apps/web/tests/expected/deepseek-messages-settings/cards.expected.md b/apps/web/tests/expected/deepseek-messages-settings/cards.expected.md
index 7df4c7ce8a..729f2b010b 100644
--- a/apps/web/tests/expected/deepseek-messages-settings/cards.expected.md
+++ b/apps/web/tests/expected/deepseek-messages-settings/cards.expected.md
@@ -8,7 +8,7 @@
   - button "打开配置文件"
   - button "关闭"
   - heading "模型" [level=2]
-  - paragraph: 填入各提供方的 API 密钥即可使用其模型。
+  - paragraph: 填入各提供商的 API 密钥即可使用其模型。
   - list:
     - listitem:
       - text: DeepSeek
@@ -43,5 +43,4 @@
           - button "添加模型"
       - button "取消"
       - button "保存"
-  - button "添加提供方"
-  - button "添加自定义提供方"
+  - button "添加模型提供商"
diff --git a/apps/web/tests/expected/models-settings-recovery/stored-error.expected.md b/apps/web/tests/expected/models-settings-recovery/stored-error.expected.md
index 69e631f4cd..84f93fc533 100644
--- a/apps/web/tests/expected/models-settings-recovery/stored-error.expected.md
+++ b/apps/web/tests/expected/models-settings-recovery/stored-error.expected.md
@@ -8,7 +8,7 @@
   - button "打开配置文件"
   - button "关闭"
   - heading "模型" [level=2]
-  - paragraph: 填入各提供方的 API 密钥即可使用其模型。
+  - paragraph: 填入各提供商的 API 密钥即可使用其模型。
   - list:
     - listitem:
       - text: openrouter
@@ -24,5 +24,4 @@
       - button "编辑 acme-gateway": 编辑
       - button "删除 acme-gateway": 删除
       - alert: "llm-pi-ai: provider \"acme-gateway\" model \"custom-model\" needs an api; the installed catalog does not describe it, so set the route's api to the wire protocol its endpoint speaks"
-  - button "添加提供方"
-  - button "添加自定义提供方"
+  - button "添加模型提供商"
diff --git a/apps/web/tests/expected/models-settings/catalog-inputs.expected.md b/apps/web/tests/expected/models-settings/catalog-inputs.expected.md
index e30e6cc2dd..d1047bfbb2 100644
--- a/apps/web/tests/expected/models-settings/catalog-inputs.expected.md
+++ b/apps/web/tests/expected/models-settings/catalog-inputs.expected.md
@@ -8,7 +8,7 @@
   - button "打开配置文件"
   - button "关闭"
   - heading "模型" [level=2]
-  - paragraph: 填入各提供方的 API 密钥即可使用其模型。
+  - paragraph: 填入各提供商的 API 密钥即可使用其模型。
   - list:
     - listitem:
       - text: minimax-cn
@@ -25,7 +25,7 @@
       - group:
         - text: 自定义设置 API 地址
         - textbox "API 地址":
-          - /placeholder: 提供方默认
+          - /placeholder: 提供商默认
         - region "模型目录":
           - text: 模型目录 已自定义模型目录
           - button "恢复默认模型"
@@ -59,5 +59,4 @@
       - text: Acme 网关 自定义
       - button "编辑 Acme 网关 (acme-gateway)": 编辑
       - button "删除 Acme 网关 (acme-gateway)": 删除
-  - button "添加提供方"
-  - button "添加自定义提供方"
+  - button "添加模型提供商"
diff --git a/apps/web/tests/expected/models-settings/configured.expected.md b/apps/web/tests/expected/models-settings/configured.expected.md
index f294ff5977..0380871342 100644
--- a/apps/web/tests/expected/models-settings/configured.expected.md
+++ b/apps/web/tests/expected/models-settings/configured.expected.md
@@ -8,7 +8,7 @@
   - button "打开配置文件"
   - button "关闭"
   - heading "模型" [level=2]
-  - paragraph: 填入各提供方的 API 密钥即可使用其模型。
+  - paragraph: 填入各提供商的 API 密钥即可使用其模型。
   - status: 已保存 minimax-cn。
   - list:
     - listitem:
@@ -16,5 +16,4 @@
       - img "API 密钥已配置"
       - button "编辑 minimax-cn": 编辑
       - button "删除 minimax-cn": 删除
-  - button "添加提供方"
-  - button "添加自定义提供方"
+  - button "添加模型提供商"
diff --git a/apps/web/tests/expected/models-settings/declared-edit.expected.md b/apps/web/tests/expected/models-settings/declared-edit.expected.md
index fcaeb76e56..6bc07d9474 100644
--- a/apps/web/tests/expected/models-settings/declared-edit.expected.md
+++ b/apps/web/tests/expected/models-settings/declared-edit.expected.md
@@ -8,7 +8,7 @@
   - button "打开配置文件"
   - button "关闭"
   - heading "模型" [level=2]
-  - paragraph: 填入各提供方的 API 密钥即可使用其模型。
+  - paragraph: 填入各提供商的 API 密钥即可使用其模型。
   - list:
     - listitem:
       - text: minimax-cn
@@ -33,9 +33,9 @@
           - text: https://gateway.acme.example/v1
         - text: API 协议
         - combobox "API 协议":
-          - option "openai-completions" [selected]
-          - option "openai-responses"
-          - option "anthropic-messages"
+          - option "OpenAI Chat Completions" [selected]
+          - option "OpenAI Responses"
+          - option "Anthropic Messages"
         - region "模型目录":
           - text: 模型目录 已自定义模型目录
           - button "恢复默认模型"
@@ -62,5 +62,4 @@
           - button "添加模型"
       - button "取消"
       - button "保存"
-  - button "添加提供方"
-  - button "添加自定义提供方"
+  - button "添加模型提供商"
diff --git a/apps/web/tests/expected/models-settings/declared.expected.md b/apps/web/tests/expected/models-settings/declared.expected.md
index d752bbb42a..269d5ebd35 100644
--- a/apps/web/tests/expected/models-settings/declared.expected.md
+++ b/apps/web/tests/expected/models-settings/declared.expected.md
@@ -8,7 +8,7 @@
   - button "打开配置文件"
   - button "关闭"
   - heading "模型" [level=2]
-  - paragraph: 填入各提供方的 API 密钥即可使用其模型。
+  - paragraph: 填入各提供商的 API 密钥即可使用其模型。
   - list:
     - listitem:
       - text: minimax-cn
@@ -19,5 +19,4 @@
       - text: Acme Gateway 自定义
       - button "编辑 Acme Gateway (acme-gateway)": 编辑
       - button "删除 Acme Gateway (acme-gateway)": 删除
-  - button "添加提供方"
-  - button "添加自定义提供方"
+  - button "添加模型提供商"
diff --git a/apps/web/tests/expected/models-settings/empty.expected.md b/apps/web/tests/expected/models-settings/empty.expected.md
index 8ce2f31098..4ddb04a13b 100644
--- a/apps/web/tests/expected/models-settings/empty.expected.md
+++ b/apps/web/tests/expected/models-settings/empty.expected.md
@@ -8,52 +8,57 @@
   - button "打开配置文件"
   - button "关闭"
   - heading "模型" [level=2]
-  - paragraph: 填入各提供方的 API 密钥即可使用其模型。
+  - paragraph: 填入各提供商的 API 密钥即可使用其模型。
   - list
-  - text: 提供方
-  - combobox "提供方":
-    - option "amazon-bedrock"
-    - option "ant-ling"
-    - option "anthropic"
-    - option "azure-openai-responses"
-    - option "baseten"
-    - option "cerebras"
-    - option "cloudflare-ai-gateway"
-    - option "cloudflare-workers-ai"
-    - option "deepseek"
-    - option "fireworks"
-    - option "github-copilot"
-    - option "google"
-    - option "google-vertex"
-    - option "groq"
-    - option "huggingface"
-    - option "kimi-coding"
-    - option "minimax"
-    - option "minimax-cn" [selected]
-    - option "mistral"
-    - option "moonshotai"
-    - option "moonshotai-cn"
-    - option "nvidia"
-    - option "openai"
-    - option "openai-codex"
-    - option "opencode"
-    - option "opencode-go"
-    - option "openrouter"
-    - option "qwen-token-plan"
-    - option "qwen-token-plan-cn"
-    - option "qwen-token-plan-individual"
-    - option "together"
-    - option "vercel-ai-gateway"
-    - option "xai"
-    - option "xiaomi"
-    - option "xiaomi-token-plan-ams"
-    - option "xiaomi-token-plan-cn"
-    - option "xiaomi-token-plan-sgp"
-    - option "zai"
-    - option "zai-coding-cn"
-  - text: API 密钥
-  - textbox "API 密钥":
-    - /placeholder: 输入 API 密钥，或留空使用环境认证
-  - group: 自定义设置
-  - button "取消"
-  - button "保存"
+  - tablist "添加方式":
+    - tab "第三方模型提供商" [selected]
+    - tab "自定义模型 API"
+  - paragraph: 从内置目录中选择 OpenAI、Anthropic、Kimi 等提供商，填入其 API 密钥即可使用。
+  - tabpanel "第三方模型提供商":
+    - text: 提供商
+    - combobox "提供商":
+      - option "amazon-bedrock"
+      - option "ant-ling"
+      - option "anthropic"
+      - option "azure-openai-responses"
+      - option "baseten"
+      - option "cerebras"
+      - option "cloudflare-ai-gateway"
+      - option "cloudflare-workers-ai"
+      - option "deepseek"
+      - option "fireworks"
+      - option "github-copilot"
+      - option "google"
+      - option "google-vertex"
+      - option "groq"
+      - option "huggingface"
+      - option "kimi-coding"
+      - option "minimax"
+      - option "minimax-cn" [selected]
+      - option "mistral"
+      - option "moonshotai"
+      - option "moonshotai-cn"
+      - option "nvidia"
+      - option "openai"
+      - option "openai-codex"
+      - option "opencode"
+      - option "opencode-go"
+      - option "openrouter"
+      - option "qwen-token-plan"
+      - option "qwen-token-plan-cn"
+      - option "qwen-token-plan-individual"
+      - option "together"
+      - option "vercel-ai-gateway"
+      - option "xai"
+      - option "xiaomi"
+      - option "xiaomi-token-plan-ams"
+      - option "xiaomi-token-plan-cn"
+      - option "xiaomi-token-plan-sgp"
+      - option "zai"
+      - option "zai-coding-cn"
+    - text: API 密钥
+    - textbox "API 密钥":
+      - /placeholder: 输入 API 密钥，或留空使用环境认证
+    - group: 自定义设置
+    - button "取消"
+    - button "保存"
diff --git a/apps/web/tests/expected/models-settings/model-picker.expected.md b/apps/web/tests/expected/models-settings/model-picker.expected.md
index e02c9b2529..01fd8e7166 100644
--- a/apps/web/tests/expected/models-settings/model-picker.expected.md
+++ b/apps/web/tests/expected/models-settings/model-picker.expected.md
@@ -1,7 +1,7 @@
 - dialog "选择要添加的模型":
   - heading "选择要添加的模型" [level=2]
   - button "关闭"
-  - paragraph: 以下是模型提供方的可用模型，勾选要添加的模型。
+  - paragraph: 以下是模型提供商的可用模型，勾选要添加的模型。
   - searchbox "搜索模型"
   - button "取消全选"
   - list:
diff --git a/apps/web/tests/expected/onboarding-deepseek-config/default-models.expected.md b/apps/web/tests/expected/onboarding-deepseek-config/default-models.expected.md
index be77e20835..c94af7233d 100644
--- a/apps/web/tests/expected/onboarding-deepseek-config/default-models.expected.md
+++ b/apps/web/tests/expected/onboarding-deepseek-config/default-models.expected.md
@@ -8,7 +8,7 @@
   - button "打开配置文件"
   - button "关闭"
   - heading "模型" [level=2]
-  - paragraph: 填入各提供方的 API 密钥即可使用其模型。
+  - paragraph: 填入各提供商的 API 密钥即可使用其模型。
   - list:
     - listitem:
       - text: DeepSeek
@@ -56,5 +56,4 @@
           - button "添加模型"
       - button "取消"
       - button "保存"
-  - button "添加提供方"
-  - button "添加自定义提供方"
+  - button "添加模型提供商"
diff --git a/apps/web/tests/expected/onboarding-deepseek-config/models.expected.md b/apps/web/tests/expected/onboarding-deepseek-config/models.expected.md
index 9e344cc148..ef3f761e7a 100644
--- a/apps/web/tests/expected/onboarding-deepseek-config/models.expected.md
+++ b/apps/web/tests/expected/onboarding-deepseek-config/models.expected.md
@@ -8,7 +8,7 @@
   - button "打开配置文件"
   - button "关闭"
   - heading "模型" [level=2]
-  - paragraph: 填入各提供方的 API 密钥即可使用其模型。
+  - paragraph: 填入各提供商的 API 密钥即可使用其模型。
   - list:
     - listitem:
       - text: DeepSeek
@@ -50,5 +50,4 @@
           - button "添加模型"
       - button "取消"
       - button "保存"
-  - button "添加提供方"
-  - button "添加自定义提供方"
+  - button "添加模型提供商"
diff --git a/apps/web/tests/expected/onboarding-usable-provider/dismissed.expected.md b/apps/web/tests/expected/onboarding-usable-provider/dismissed.expected.md
index 0279600e91..f955841948 100644
--- a/apps/web/tests/expected/onboarding-usable-provider/dismissed.expected.md
+++ b/apps/web/tests/expected/onboarding-usable-provider/dismissed.expected.md
@@ -8,56 +8,61 @@
   - button "打开配置文件"
   - button "关闭"
   - heading "模型" [level=2]
-  - paragraph: 填入各提供方的 API 密钥即可使用其模型。
+  - paragraph: 填入各提供商的 API 密钥即可使用其模型。
   - list:
     - listitem:
       - text: DeepSeek
       - img "API 密钥缺失"
       - button "编辑 DeepSeek (deepseek-official)": 编辑
-  - text: 提供方
-  - combobox "提供方":
-    - option "amazon-bedrock"
-    - option "ant-ling"
-    - option "anthropic"
-    - option "azure-openai-responses"
-    - option "baseten"
-    - option "cerebras"
-    - option "cloudflare-ai-gateway"
-    - option "cloudflare-workers-ai"
-    - option "deepseek"
-    - option "fireworks"
-    - option "github-copilot"
-    - option "google"
-    - option "google-vertex"
-    - option "groq"
-    - option "huggingface"
-    - option "kimi-coding"
-    - option "minimax"
-    - option "minimax-cn" [selected]
-    - option "mistral"
-    - option "moonshotai"
-    - option "moonshotai-cn"
-    - option "nvidia"
-    - option "openai"
-    - option "openai-codex"
-    - option "opencode"
-    - option "opencode-go"
-    - option "openrouter"
-    - option "qwen-token-plan"
-    - option "qwen-token-plan-cn"
-    - option "qwen-token-plan-individual"
-    - option "together"
-    - option "vercel-ai-gateway"
-    - option "xai"
-    - option "xiaomi"
-    - option "xiaomi-token-plan-ams"
-    - option "xiaomi-token-plan-cn"
-    - option "xiaomi-token-plan-sgp"
-    - option "zai"
-    - option "zai-coding-cn"
-  - text: API 密钥
-  - textbox "API 密钥":
-    - /placeholder: 输入 API 密钥，或留空使用环境认证
-  - group: 自定义设置
-  - button "取消"
-  - button "保存"
+  - tablist "添加方式":
+    - tab "第三方模型提供商" [selected]
+    - tab "自定义模型 API"
+  - paragraph: 从内置目录中选择 OpenAI、Anthropic、Kimi 等提供商，填入其 API 密钥即可使用。
+  - tabpanel "第三方模型提供商":
+    - text: 提供商
+    - combobox "提供商":
+      - option "amazon-bedrock"
+      - option "ant-ling"
+      - option "anthropic"
+      - option "azure-openai-responses"
+      - option "baseten"
+      - option "cerebras"
+      - option "cloudflare-ai-gateway"
+      - option "cloudflare-workers-ai"
+      - option "deepseek"
+      - option "fireworks"
+      - option "github-copilot"
+      - option "google"
+      - option "google-vertex"
+      - option "groq"
+      - option "huggingface"
+      - option "kimi-coding"
+      - option "minimax"
+      - option "minimax-cn" [selected]
+      - option "mistral"
+      - option "moonshotai"
+      - option "moonshotai-cn"
+      - option "nvidia"
+      - option "openai"
+      - option "openai-codex"
+      - option "opencode"
+      - option "opencode-go"
+      - option "openrouter"
+      - option "qwen-token-plan"
+      - option "qwen-token-plan-cn"
+      - option "qwen-token-plan-individual"
+      - option "together"
+      - option "vercel-ai-gateway"
+      - option "xai"
+      - option "xiaomi"
+      - option "xiaomi-token-plan-ams"
+      - option "xiaomi-token-plan-cn"
+      - option "xiaomi-token-plan-sgp"
+      - option "zai"
+      - option "zai-coding-cn"
+    - text: API 密钥
+    - textbox "API 密钥":
+      - /placeholder: 输入 API 密钥，或留空使用环境认证
+    - group: 自定义设置
+    - button "取消"
+    - button "保存"
diff --git a/apps/web/tests/models-settings-recovery.e2e.ts b/apps/web/tests/models-settings-recovery.e2e.ts
index 1087f8be3d..3e2d664b93 100644
--- a/apps/web/tests/models-settings-recovery.e2e.ts
+++ b/apps/web/tests/models-settings-recovery.e2e.ts
@@ -54,19 +54,21 @@ describe('web e2e: repairs a stored provider after catalog drift', () => {
     }
   })
 
-  it('shows the failed provider beside healthy providers and keeps both add actions usable', async () => {
+  it('shows the failed provider beside healthy providers and keeps the add action usable', async () => {
     onTestFailed(() => saveFailureShot(page, 'models-settings-recovery'))
     const dialog = page.getByRole('dialog', { name: '设置' })
     expect(await dialog.getByRole('button', { name: '编辑 openrouter', exact: true }).count()).toBe(1)
     expect(await dialog.getByRole('button', { name: '编辑 zai', exact: true }).count()).toBe(1)
     expect(await dialog.getByRole('button', { name: '编辑 acme-gateway', exact: true }).count()).toBe(1)
     expect(await dialog.getByText(CUSTOM_FAILURE, { exact: true }).count()).toBe(1)
-    expect(await dialog.getByRole('button', { name: '添加提供方', exact: true }).isEnabled()).toBe(true)
-    expect(await dialog.getByRole('button', { name: '添加自定义提供方', exact: true }).isEnabled()).toBe(true)
+    expect(await dialog.getByRole('button', { name: '添加模型提供商', exact: true }).isEnabled()).toBe(true)
     await compareOrRefreshGolden(EXPECTED, await captureStableAria(page, '[role="dialog"]', scaffold.workspaceCwd), webSnapshotMode())
 
-    await dialog.getByRole('button', { name: '添加提供方', exact: true }).click()
-    await dialog.getByLabel('提供方', { exact: true }).selectOption('minimax-cn')
+    await dialog.getByRole('button', { name: '添加模型提供商', exact: true }).click()
+    // Both modes stay offered beside a failed route: the card opens on the
+    // third-party mode with the custom-API segment enabled.
+    expect(await dialog.getByRole('tab', { name: '自定义模型 API' }).isEnabled()).toBe(true)
+    await dialog.getByLabel('提供商', { exact: true }).selectOption('minimax-cn')
     await dialog.getByRole('button', { name: '保存', exact: true }).click()
     await dialog.getByText('已保存 minimax-cn。', { exact: true }).waitFor()
     expect(await readFile(join(home, 'settings.yaml'), 'utf8')).toContain('minimax-cn: {}')
diff --git a/apps/web/tests/models-settings.e2e.ts b/apps/web/tests/models-settings.e2e.ts
index 888b5d5d54..6d514cf97b 100644
--- a/apps/web/tests/models-settings.e2e.ts
+++ b/apps/web/tests/models-settings.e2e.ts
@@ -1,5 +1,6 @@
 // Web e2e scenario: the Models settings page end to end through the real
-// wire — the add card offers the dormant pi-ai catalog, a blank key saves a
+// wire — the one add card offers the dormant pi-ai catalog on its third-party
+// mode and the create form on its custom-API mode, a blank key saves a
 // reference-free profile for provider-native auth, and typing an API key later
 // stores it write-only under the derived reference (`MINIMAX_CN_API_KEY`)
 // while the settings document records only that reference. Each saved row
@@ -75,15 +76,19 @@ describe('web e2e: Models settings page configures a dormant provider', () => {
     const dialog = page.getByRole('dialog', { name: '设置' })
     await dialog.waitFor({ timeout: 10_000 })
     await dialog.getByRole('button', { name: '模型' }).click()
-    await dialog.getByText('填入各提供方的 API 密钥即可使用其模型。').waitFor({ timeout: 10_000 })
+    await dialog.getByText('填入各提供商的 API 密钥即可使用其模型。').waitFor({ timeout: 10_000 })
     // The dormant pi-ai adapter contributes its whole installed catalog; no
     // provider is configured yet, so the page is one add button.
-    const add = dialog.getByRole('button', { name: '添加提供方' })
+    const add = dialog.getByRole('button', { name: '添加模型提供商' })
     await add.waitFor({ timeout: 10_000 })
     // The button enables once the dormant catalog lands in the join.
     await expect.poll(async () => add.isEnabled(), { timeout: 10_000 }).toBe(true)
     await add.click()
-    const pick = dialog.getByLabel('提供方')
+    // The card opens on the third-party mode, whose panel carries the select.
+    const modes = dialog.getByRole('tablist', { name: '添加方式' })
+    await modes.waitFor({ timeout: 10_000 })
+    expect(await modes.getByRole('tab', { name: '第三方模型提供商' }).getAttribute('aria-selected')).toBe('true')
+    const pick = dialog.getByLabel('提供商', { exact: true })
     await pick.waitFor({ timeout: 10_000 })
     await expect.poll(async () => pick.locator('option').count(), { timeout: 10_000 }).toBeGreaterThan(30)
     const options = await pick.locator('option').allTextContents()
@@ -239,22 +244,29 @@ describe('web e2e: Models settings page configures a dormant provider', () => {
   it('declares a route the adapter does not ship', async () => {
     onTestFailed(() => saveFailureShot(page, 'web-e2e-models-declare'))
     const dialog = page.getByRole('dialog', { name: '设置' })
-    const declare = dialog.getByRole('button', { name: '添加自定义提供方' })
-    await expect.poll(async () => declare.isEnabled(), { timeout: 10_000 }).toBe(true)
-    await declare.click()
-    await dialog.getByLabel('Provider ID').fill('acme-gateway')
-    await dialog.getByLabel('显示名称').fill('Acme Gateway')
-    await dialog.getByLabel('API 地址').fill('https://gateway.acme.example/v1')
+    const add = dialog.getByRole('button', { name: '添加模型提供商' })
+    await expect.poll(async () => add.isEnabled(), { timeout: 10_000 }).toBe(true)
+    await add.click()
+    // The custom-API mode is the second segment of the one add card; its
+    // purpose line replaces the catalog one.
+    await dialog.getByRole('tab', { name: '自定义模型 API' }).click()
+    await dialog.getByText('连接中转站、自部署服务或其他兼容 OpenAI / Anthropic 协议的接口，需填写 API 地址、协议和模型。').waitFor({ timeout: 10_000 })
+    // The third-party panel stays mounted, hidden, beside this one, so the
+    // form is addressed through its own panel.
+    const custom = dialog.getByRole('tabpanel', { name: '自定义模型 API' })
+    await custom.getByLabel('Provider ID').fill('acme-gateway')
+    await custom.getByLabel('显示名称').fill('Acme Gateway')
+    await custom.getByLabel('API 地址').fill('https://gateway.acme.example/v1')
     // No reasoning effort on a provider card at all: effort is a per-model
     // capability, the models under one provider disagree about it, and a
     // switch in the composer already records provider+model+effort together.
     expect(await dialog.getByLabel('推理强度').count()).toBe(0)
-    await dialog.getByRole('button', { name: '添加模型' }).click()
-    await dialog.getByLabel('模型 ID 1').fill('acme-large')
-    await dialog.getByRole('button', { name: '模型选项 1' }).click()
-    expect(await dialog.getByRole('group', { name: '输入类型 1' }).getByRole('checkbox', { name: '图片' }).isChecked()).toBe(false)
-    await dialog.getByRole('group', { name: '输入类型 1' }).getByRole('checkbox', { name: '图片' }).check()
-    await dialog.getByRole('button', { name: '创建提供方', exact: true }).click()
+    await custom.getByRole('button', { name: '添加模型', exact: true }).click()
+    await custom.getByLabel('模型 ID 1').fill('acme-large')
+    await custom.getByRole('button', { name: '模型选项 1' }).click()
+    expect(await custom.getByRole('group', { name: '输入类型 1' }).getByRole('checkbox', { name: '图片' }).isChecked()).toBe(false)
+    await custom.getByRole('group', { name: '输入类型 1' }).getByRole('checkbox', { name: '图片' }).check()
+    await custom.getByRole('button', { name: '创建提供商', exact: true }).click()
 
     const row = dialog.getByText('Acme Gateway', { exact: true }).first()
     await row.waitFor({ timeout: 10_000 })
diff --git a/apps/web/tests/onboarding-deepseek-config.e2e.ts b/apps/web/tests/onboarding-deepseek-config.e2e.ts
index 96ddb5e24c..d8e201a49d 100644
--- a/apps/web/tests/onboarding-deepseek-config.e2e.ts
+++ b/apps/web/tests/onboarding-deepseek-config.e2e.ts
@@ -233,7 +233,7 @@ describe.skipIf(MODE === 'record')('web e2e: first-run DeepSeek credential setup
     for (let index = 0; index < 2; index++) {
       await settings.getByRole('button', { name: /删除模型/ }).first().click()
     }
-    await settings.getByRole('button', { name: '添加模型' }).click()
+    await settings.getByRole('button', { name: '添加模型', exact: true }).click()
     const customModelId = settings.getByLabel('模型 ID 1')
     await customModelId.fill('private-preview')
     await settings.getByLabel('显示名称 1').fill('Private Preview')
diff --git a/apps/web/tests/onboarding-usable-provider.e2e.ts b/apps/web/tests/onboarding-usable-provider.e2e.ts
index d16ed51822..efc04dcdaa 100644
--- a/apps/web/tests/onboarding-usable-provider.e2e.ts
+++ b/apps/web/tests/onboarding-usable-provider.e2e.ts
@@ -58,10 +58,10 @@ describe.skipIf(MODE === 'record')('web e2e: another usable provider ends first-
     const setupKey = settings.getByRole('textbox', { name: 'API 密钥', exact: true })
     await setupKey.waitFor({ timeout: 10_000 })
 
-    const add = settings.getByRole('button', { name: '添加提供方' })
+    const add = settings.getByRole('button', { name: '添加模型提供商' })
     await expect.poll(async () => add.isEnabled(), { timeout: 10_000 }).toBe(true)
     await add.click()
-    const pick = settings.getByLabel('提供方')
+    const pick = settings.getByLabel('提供商', { exact: true })
     await pick.waitFor({ timeout: 10_000 })
     await pick.selectOption('minimax-cn')
     await expect.poll(
@@ -72,7 +72,7 @@ describe.skipIf(MODE === 'record')('web e2e: another usable provider ends first-
     // Cancelling the setup card must not close the independent add-provider
     // draft beside it.
     await settings.getByRole('button', { name: '取消', exact: true }).first().click()
-    expect(await settings.getByLabel('提供方').count()).toBe(1)
+    expect(await settings.getByLabel('提供商', { exact: true }).count()).toBe(1)
     await expect.poll(
       async () => settings.getByRole('textbox', { name: 'API 密钥', exact: true }).count(),
       { timeout: 10_000 },
diff --git a/docs/user/guide/providers-custom-form.png b/docs/user/guide/providers-custom-form.png
index bbbedde794..82ca074212 100644
Binary files a/docs/user/guide/providers-custom-form.png and b/docs/user/guide/providers-custom-form.png differ
diff --git a/docs/user/guide/providers-custom-form.zh.png b/docs/user/guide/providers-custom-form.zh.png
index 2c48120331..9a2b7661ca 100644
Binary files a/docs/user/guide/providers-custom-form.zh.png and b/docs/user/guide/providers-custom-form.zh.png differ
diff --git a/docs/user/guide/providers-models-page.png b/docs/user/guide/providers-models-page.png
index f3ffe6e90f..fb87367d54 100644
Binary files a/docs/user/guide/providers-models-page.png and b/docs/user/guide/providers-models-page.png differ
diff --git a/docs/user/guide/providers-models-page.zh.png b/docs/user/guide/providers-models-page.zh.png
index b5d1f4597c..694f1b6f16 100644
Binary files a/docs/user/guide/providers-models-page.zh.png and b/docs/user/guide/providers-models-page.zh.png differ
diff --git a/docs/user/guide/providers.i18n.yaml b/docs/user/guide/providers.i18n.yaml
index 3097c5ae9e..f0043c374d 100644
--- a/docs/user/guide/providers.i18n.yaml
+++ b/docs/user/guide/providers.i18n.yaml
@@ -2,5 +2,5 @@
 # side as of the last confirmed-consistent state. Both languages carry equal authority;
 # after editing either side, bring the other along and re-record with:
 #   pnpm run verify-translation-pairing --write docs/user/guide/providers.md
-providers.md: e578ff96a373c96ebe646126d6c60d7bcc62bbc1
-providers.zh.md: fd0859f0b235efb5fe8ecc3c248a2ac2bbaeb495
+providers.md: ef016ef39d5e34f3ce3dd2feb6213b9597668592
+providers.zh.md: 15fd229ce222a020e897f08c15afcd843791a3bb
diff --git a/docs/user/guide/providers.md b/docs/user/guide/providers.md
index e578ff96a3..ef016ef39d 100644
--- a/docs/user/guide/providers.md
+++ b/docs/user/guide/providers.md
@@ -8,21 +8,21 @@ This guide assumes you started the Web UI through the [root README](../../../REA
 
 Open **Settings → Models**. The DeepSeek card exposes one API-key field; enter the key and save it.
 
-![The Models page: the DeepSeek card, with Add provider and Add a custom provider below it](providers-models-page.png)
+![The Models page: the DeepSeek card, with Add model provider below it](providers-models-page.png)
 
 Keys are write-only. The page receives a redacted descriptor after saving, never the literal secret. The key is stored in `$DSH_HOME/.credentials.yaml`, while settings retain only its credential reference.
 
-## Add a built-in provider
+## Add a third-party provider
 
-Choose **Add provider** and pick a provider dsh ships with; the list shows provider ids such as `anthropic`, `openai`, `moonshotai` for Kimi, or `zai` for GLM. Enter its API key and save. The installed catalog supplies the endpoint, protocol, and model list.
+Choose **Add model provider**. The card opens on **Third-party model provider**: pick a provider dsh ships with — the list shows provider ids such as `anthropic`, `openai`, `moonshotai` for Kimi, or `zai` for GLM — enter its API key, and save. The installed catalog supplies the endpoint, protocol, and model list.
 
 Providers that sign in with OAuth, such as Codex, are not supported here yet.
 
-## Add a custom provider
+## Add a custom model API
 
-Choose **Add a custom provider** for a company gateway, self-hosted server, or provider absent from the installed catalog. Supply a lowercase Provider ID, base URL, API protocol, credential, and at least one model. The **API protocol** must be the one your gateway speaks, and the form offers three: `openai-completions` for OpenAI Chat Completions, `openai-responses` for the OpenAI Responses API, and `anthropic-messages` for the Anthropic Messages API. A provider speaks one protocol, so a gateway that serves two needs two providers.
+Switch the card to **Custom model API** for a relay, a company gateway, a self-hosted server, or any provider absent from the installed catalog. Supply a lowercase Provider ID, base URL, API protocol, credential, and at least one model. The **API protocol** must be the one your gateway speaks, and the picker offers three: OpenAI Chat Completions, OpenAI Responses, and Anthropic Messages, stored in `settings.yaml` as `openai-completions`, `openai-responses`, and `anthropic-messages`. A provider speaks one protocol, so a gateway that serves two needs two providers.
 
-![The custom provider form: Provider ID, display name, base URL, API protocol, and API key](providers-custom-form.png)
+![The custom model API form: Provider ID, display name, base URL, API protocol, and API key](providers-custom-form.png)
 
 The Provider ID is permanent because requests, saved sessions, model defaults, and credential references use it. To rename a provider, add a new provider and delete the old one. The display name, base URL, protocol, credential, and models remain editable.
 
diff --git a/docs/user/guide/providers.zh.md b/docs/user/guide/providers.zh.md
index fd0859f0b2..15fd229ce2 100644
--- a/docs/user/guide/providers.zh.md
+++ b/docs/user/guide/providers.zh.md
@@ -8,39 +8,39 @@
 
 打开**设置 → 模型**。DeepSeek 卡片提供一个 API 密钥字段；输入密钥并保存。
 
-![模型页：DeepSeek 卡片，以及添加提供方与添加自定义提供方两个入口](providers-models-page.zh.png)
+![模型页：DeepSeek 卡片，及其下方的添加模型提供商入口](providers-models-page.zh.png)
 
 密钥是只写的。保存后，页面只会收到脱敏描述符，永远不会收到明文密钥。密钥存储在 `$DSH_HOME/.credentials.yaml` 中，settings 只保留它的凭据引用。
 
-## 添加内置提供方
+## 添加第三方提供商
 
-选择**添加提供方**，选取 dsh 自带的提供方；列表显示的是提供方 id，例如 `anthropic`、`openai`、Kimi 对应的 `moonshotai`、GLM 对应的 `zai`。输入其 API 密钥并保存。已安装目录会提供端点、协议和模型列表。
+选择**添加模型提供商**。卡片默认打开在**第三方模型提供商**：选取 dsh 自带的提供商——列表显示的是提供商 id，例如 `anthropic`、`openai`、Kimi 对应的 `moonshotai`、GLM 对应的 `zai`——输入其 API 密钥并保存。已安装目录会提供端点、协议和模型列表。
 
-通过 OAuth 登录的提供方（例如 Codex）暂不支持。
+通过 OAuth 登录的提供商（例如 Codex）暂不支持。
 
-## 添加自定义提供方
+## 添加自定义模型 API
 
-对于公司网关、自建服务器或已安装目录中不存在的提供方，选择**添加自定义提供方**。提供小写 Provider ID、基础 URL、API 协议、凭据和至少一个模型。**API 协议**必须选网关实际使用的那一种，表单提供三种：`openai-completions` 对应 OpenAI Chat Completions，`openai-responses` 对应 OpenAI Responses API，`anthropic-messages` 对应 Anthropic Messages API。一个提供方只使用一种协议，网关同时提供两种时需要建两个提供方。
+对于中转站、公司网关、自建服务器或已安装目录中不存在的提供商，把卡片切换到**自定义模型 API**。提供小写 Provider ID、基础 URL、API 协议、凭据和至少一个模型。**API 协议**必须选网关实际使用的那一种，选择框提供三种：OpenAI Chat Completions、OpenAI Responses 和 Anthropic Messages，在 `settings.yaml` 中分别存为 `openai-completions`、`openai-responses` 和 `anthropic-messages`。一个提供商只使用一种协议，网关同时提供两种时需要建两个提供商。
 
-![自定义提供方表单：Provider ID、显示名称、API 地址、API 协议、API 密钥](providers-custom-form.zh.png)
+![自定义模型 API 表单：Provider ID、显示名称、API 地址、API 协议、API 密钥](providers-custom-form.zh.png)
 
-Provider ID 是永久的，因为请求、已保存会话、模型默认值和凭据引用都会使用它。如需重命名提供方，请添加新提供方并删除旧提供方。显示名称、基础 URL、协议、凭据和模型仍可编辑。
+Provider ID 是永久的，因为请求、已保存会话、模型默认值和凭据引用都会使用它。如需重命名提供商，请添加新提供商并删除旧提供商。显示名称、基础 URL、协议、凭据和模型仍可编辑。
 
 ### 探测模型
 
-在**模型目录**中选择**获取可用模型**，即可询问端点它提供哪些模型。请求使用表单当前的 API 地址、协议和密钥，已保存的提供方则用已存储的密钥；响应会打开一个可搜索的选择框，搜索、勾选想要的模型，再点**添加所选**。保存或创建提供方之前不会存储任何内容。
+在**模型目录**中选择**获取可用模型**，即可询问端点它提供哪些模型。请求使用表单当前的 API 地址、协议和密钥，已保存的提供商则用已存储的密钥；响应会打开一个可搜索的选择框，搜索、勾选想要的模型，再点**添加所选**。保存或创建提供商之前不会存储任何内容。
 
-探测读取的是常见网关公开的列表格式，但并非每个端点都用这些格式作答，所以它只是便利手段而非保证：探测失败或列表为空时，手动添加模型 ID 即可，效果完全一样。内置提供方一律由已安装目录作答，即使其 API 地址指向网关也是如此，要查看网关实际提供的模型，请通过自定义提供方探测。
+探测读取的是常见网关公开的列表格式，但并非每个端点都用这些格式作答，所以它只是便利手段而非保证：探测失败或列表为空时，手动添加模型 ID 即可，效果完全一样。内置提供商一律由已安装目录作答，即使其 API 地址指向网关也是如此，要查看网关实际提供的模型，请通过自定义提供商探测。
 
 ## 选择模型
 
-已配置的提供方会出现在模型选择器中。选择模型也会将其设为新会话的默认值。已发送过请求的会话会保留自身日志中记录的模型。
+已配置的提供商会出现在模型选择器中。选择模型也会将其设为新会话的默认值。已发送过请求的会话会保留自身日志中记录的模型。
 
-如果已保存默认值指向已删除的提供方，输入框会显示**选择模型**，并在选择其他模型前阻止输入。
+如果已保存默认值指向已删除的提供商，输入框会显示**选择模型**，并在选择其他模型前阻止输入。
 
 ## 进阶配置
 
-自动生成的[插件配置目录](../../config-catalog.zh.md)列出每个插件的所有受支持字段与默认值；[`dsh-llm-pi-ai`](../../config-catalog.zh.md#deepseek-aidsh-llm-pi-ai) 就是本页所配置的那个提供方段落。[`dsh-llm-pi-ai`](../../../packages/llm/llm-pi-ai/README.zh.md) 和 [`dsh-llm-deepseek`](../../../packages/llm/llm-deepseek/README.zh.md) 参考文档负责直接 `settings.yaml` 配置、目录解析、推理控制、凭据与适配器错误。
+自动生成的[插件配置目录](../../config-catalog.zh.md)列出每个插件的所有受支持字段与默认值；[`dsh-llm-pi-ai`](../../config-catalog.zh.md#deepseek-aidsh-llm-pi-ai) 就是本页所配置的那个提供商段落。[`dsh-llm-pi-ai`](../../../packages/llm/llm-pi-ai/README.zh.md) 和 [`dsh-llm-deepseek`](../../../packages/llm/llm-deepseek/README.zh.md) 参考文档负责直接 `settings.yaml` 配置、目录解析、推理控制、凭据与适配器错误。
 
 ::: tip 其他设置
 模型页提供 API 密钥、显示名称、API 地址、API 协议，以及每个模型的 ID、显示名称、上下文窗口、最大输出 token 数和输入类型。推理等级、请求兼容性开关、请求头、超时和重试策略在 `$DSH_HOME/settings.yaml` 中设置，也就是模型页写入的同一份文档。可以直接编辑它；浏览器与服务器在同一台机器时，也可以点击设置页顶部的**打开配置文件**打开它。适配器会在下一次请求时重新读取，无需重启任何东西。下面各小节介绍多数网关会用到的字段。
@@ -48,9 +48,9 @@ Provider ID 是永久的，因为请求、已保存会话、模型默认值和
 
 ### 图片输入
 
-在**设置 → 模型**中编辑提供方，打开**自定义设置**并展开该模型的**模型选项**。**输入类型**独占容量字段下方的一行。对于支持图片的模型，勾选**图片**并保存。没有继承图片能力的新自定义模型默认勾选**文本**。至少保留一种输入类型；仅图片模型需先勾选图片，再取消文本。
+在**设置 → 模型**中编辑提供商，打开**自定义设置**并展开该模型的**模型选项**。**输入类型**独占容量字段下方的一行。对于支持图片的模型，勾选**图片**并保存。没有继承图片能力的新自定义模型默认勾选**文本**。至少保留一种输入类型；仅图片模型需先勾选图片，再取消文本。
 
-复选框将 pi-ai 模型的选择保存为 `input`，将直连 DeepSeek 适配器的选择保存为 `inputModalities`。也可以在 `$DSH_HOME/settings.yaml` 中编辑模型；例如，以下自定义 pi-ai 提供方声明了一个纯文本模型和一个视觉模型：
+复选框将 pi-ai 模型的选择保存为 `input`，将直连 DeepSeek 适配器的选择保存为 `inputModalities`。也可以在 `$DSH_HOME/settings.yaml` 中编辑模型；例如，以下自定义 pi-ai 提供商声明了一个纯文本模型和一个视觉模型：
 
 ```yaml
 llm-pi-ai:
@@ -86,7 +86,7 @@ llm-pi-ai:
         - id: second-model
 ```
 
-`defaultInput` 是回退值而不是覆盖值，默认为 `[text]`：在内置提供方上，它只为其目录未描述的模型作答，因此绝不会把目录中本就具备图片能力的模型的该能力去掉。要收窄这类模型，请用它自己的 `input`。内置提供方没有显式 `models` 列表时，写在 `modelOverrides` 下，以模型 id 为键：
+`defaultInput` 是回退值而不是覆盖值，默认为 `[text]`：在内置提供商上，它只为其目录未描述的模型作答，因此绝不会把目录中本就具备图片能力的模型的该能力去掉。要收窄这类模型，请用它自己的 `input`。内置提供商没有显式 `models` 列表时，写在 `modelOverrides` 下，以模型 id 为键：
 
 ```yaml
 llm-pi-ai:
@@ -99,11 +99,11 @@ llm-pi-ai:
 
 在 pi-ai 配置中，除模型自身的 `input` 外，每个列表都至少要写一项模态；模型自身的空列表与省略它同义。未知模态在任何位置写入都会被拒绝。
 
-这两个字段都是对你端点的断言，而不是对它的检查。声明了端点并不提供的图片能力的模型不会在这里被拦下，改由提供方拒绝该请求。
+这两个字段都是对你端点的断言，而不是对它的检查。声明了端点并不提供的图片能力的模型不会在这里被拦下，改由提供商拒绝该请求。
 
 ### 推理等级
 
-对于声明了推理等级的模型，模型选择器会提供**推理等级**菜单。内置提供方的模型从已安装目录继承其等级。手动录入的模型不声明任何等级，因此模型菜单里不会出现推理等级项，由端点自身的默认值决定模型是否思考。请在 `$DSH_HOME/settings.yaml` 中用 `reasoningEfforts` 声明等级：
+对于声明了推理等级的模型，模型选择器会提供**推理等级**菜单。内置提供商的模型从已安装目录继承其等级。手动录入的模型不声明任何等级，因此模型菜单里不会出现推理等级项，由端点自身的默认值决定模型是否思考。请在 `$DSH_HOME/settings.yaml` 中用 `reasoningEfforts` 声明等级：
 
 ```yaml
 llm-pi-ai:
@@ -136,7 +136,7 @@ llm-pi-ai:
             max: max
 ```
 
-网关并不提供推理能力的内置提供方模型，可在 `modelOverrides` 下用 `reasoningEfforts: false` 去掉其等级；之后再为它选择等级会被拒绝并报 `UNSUPPORTED_REASONING_EFFORT`。DeepSeek 自身的路由不需要以上任何配置：其模型已经提供 `off`、`low`、`high` 和 `max`，`llm-deepseek.reasoningEffort` 设置选择器的起始默认值：
+网关并不提供推理能力的内置提供商模型，可在 `modelOverrides` 下用 `reasoningEfforts: false` 去掉其等级；之后再为它选择等级会被拒绝并报 `UNSUPPORTED_REASONING_EFFORT`。DeepSeek 自身的路由不需要以上任何配置：其模型已经提供 `off`、`low`、`high` 和 `max`，`llm-deepseek.reasoningEffort` 设置选择器的起始默认值：
 
 ```yaml
 llm-deepseek:
@@ -181,8 +181,8 @@ llm-pi-ai:
 
 ## 排错
 
-- **`MISSING_CREDENTIAL`**：通过模型页存储提供方密钥，或提供被引用的环境变量。
-- **`UNKNOWN_MODEL`**：选择已配置的模型，或向自定义提供方添加缺失的模型。
+- **`MISSING_CREDENTIAL`**：通过模型页存储提供商密钥，或提供被引用的环境变量。
+- **`UNKNOWN_MODEL`**：选择已配置的模型，或向自定义提供商添加缺失的模型。
 - **获取可用模型返回 401**：检查密钥。模型发现会调用 OpenAI 兼容的 `GET /models` 端点；对于不提供该端点的服务，请手动输入模型。
 - **获取可用模型提示既没有 `data` 数组也没有 `models` 对象**：端点返回的列表格式不在探测的读取范围内。请手动输入模型。
 - **密钥与地址都正确，网关却拒绝每一个请求**：它的请求形状与 OpenAI 不同。先在路由上设 `compat.supportsDeveloperRole: false` 与 `compat.maxTokensField: max_tokens`。
@@ -190,5 +190,5 @@ llm-pi-ai:
 - **手动录入的模型没有推理等级菜单**：该模型没有声明任何等级。在 `settings.yaml` 中给该模型加上 `reasoningEfforts`。
 - **`off` 无法让 DeepSeek 模型停止思考**：留空的 `off` 不发送任何推理字段，默认思考的端点就继续思考。请在模型或路由上设置 `compat.thinkingFormat: deepseek`。
 - **某个 compat 开关因没有值而被拒绝**：冒号后什么都没写。给它一个值，或删掉该键以沿用已安装 catalog 的值。
-- **图片在发送前被拒绝**：该模型未声明图片模态。请给自定义提供方的模型加上 `input: [text, image]`；在 DeepSeek 自身的路由上，请从配置的目录中选择支持图片的条目（默认为 `deepseek-flash`），并确认网关提供该模型且支持图片输入。
-- **提供方拒绝了带图片的请求**：该模型声明了其端点实际并不提供的图片能力。请从授予它图片能力的那个列表中移除 `image`——可能是模型的 `input`，也可能是路由的 `defaultInput`——然后开启新会话：附加的图片会留在会话日志里，因此在会话离开它之前，同一个请求会不断重复。
+- **图片在发送前被拒绝**：该模型未声明图片模态。请给自定义提供商的模型加上 `input: [text, image]`；在 DeepSeek 自身的路由上，请从配置的目录中选择支持图片的条目（默认为 `deepseek-flash`），并确认网关提供该模型且支持图片输入。
+- **提供商拒绝了带图片的请求**：该模型声明了其端点实际并不提供的图片能力。请从授予它图片能力的那个列表中移除 `image`——可能是模型的 `input`，也可能是路由的 `defaultInput`——然后开启新会话：附加的图片会留在会话日志里，因此在会话离开它之前，同一个请求会不断重复。
diff --git a/packages/client/ui-settings-models/README.i18n.yaml b/packages/client/ui-settings-models/README.i18n.yaml
index c3c33c6f90..c81192e0e4 100644
--- a/packages/client/ui-settings-models/README.i18n.yaml
+++ b/packages/client/ui-settings-models/README.i18n.yaml
@@ -2,5 +2,5 @@
 # side as of the last confirmed-consistent state. Both languages carry equal authority;
 # after editing either side, bring the other along and re-record with:
 #   pnpm run verify-translation-pairing --write packages/client/ui-settings-models/README.md
-README.md: c1567814b5de39f9b8819fcc8848e36c0a6ff842
-README.zh.md: baa0f02e8274a7146f9eade36d3ad464ac944b27
+README.md: 619941e125d6c8607246b5924c66c9a11af47948
+README.zh.md: c4ba114dbb6501f564f810efc339114ea902604e
diff --git a/packages/client/ui-settings-models/README.md b/packages/client/ui-settings-models/README.md
index c1567814b5..619941e125 100644
--- a/packages/client/ui-settings-models/README.md
+++ b/packages/client/ui-settings-models/README.md
@@ -43,7 +43,7 @@ Expand **Customized settings → Model options** to edit each model. Both provid
 
 ### Adding and deleting providers
 
-The add flow is a card carrying the dormant-directory provider select — a bare-mounted `llm-pi-ai` offers its whole installed catalog before any route exists. **Add a custom provider** declares a route pi-ai does not ship; the create card asks for a unique **Provider ID**, an endpoint, a protocol, and at least one uniquely-identified model, because nothing can default those. The endpoint placeholder follows the selected protocol: OpenAI Chat Completions and Responses show `https://gateway.example/v1`, while Anthropic Messages shows `https://gateway.example` because its SDK appends `/v1/messages`. Changing the protocol preserves any address already entered. The endpoint must be a parseable HTTP or HTTPS URL; localhost, IPv4 and IPv6 literals, and custom ports remain valid. A syntax error blocks both discovery and creation at the field, while a request failure remains a separate provider error. **Fetch available models** asks the `llm/discoverModels` Remote about the endpoint the form shows, so adding a provider is one pass instead of save-then-return; the reply opens a searchable picker rather than being written, and nothing is written until **Add selected**. Each selected candidate copies its id, display name, context window, output-token cap, and disclosed input types into the editable row when disclosed, while an existing row retains its user-tuned values. Search matches model ids and optional display names without clearing hidden selections. **Select all** adds the visible results, while **Deselect all** clears the entire selection so hidden results cannot be adopted accidentally. A row is deletable only when the user layer alone carries it (removal restores the composition base), and its confirmation dialog names the provider.
+**Add model provider** is the one entry to the add card, offered while any editor-backed settings namespace is mounted and enabled while either mode below can proceed. The card opens on a segmented mode switch with a one-line purpose beneath it: **Third-party model provider** carries the dormant-directory provider select over the provider editor — a bare-mounted `llm-pi-ai` offers its whole installed catalog before any route exists — and **Custom model API** carries the create form for a route pi-ai does not ship: a relay, a self-hosted server, or any other OpenAI- or Anthropic-compatible endpoint. A mode is offered only while its namespace is mounted; with one mode the card shows that form alone, and a mode whose namespace has nothing left to adopt or no protocol to declare is disabled. A panel mounts the first time its mode is shown and stays mounted, hidden, while the card is open, so switching modes discards neither draft. The custom form asks for a unique **Provider ID**, an endpoint, a protocol, and at least one uniquely-identified model, because nothing can default those. The protocol picker names each protocol by its product name — OpenAI Chat Completions, OpenAI Responses, Anthropic Messages — while the stored value stays the schema identifier; a protocol the adapter adds before this page names it shows that identifier. The endpoint placeholder follows the selected protocol: OpenAI Chat Completions and Responses show `https://gateway.example/v1`, while Anthropic Messages shows `https://gateway.example` because its SDK appends `/v1/messages`. Changing the protocol preserves any address already entered. The endpoint must be a parseable HTTP or HTTPS URL; localhost, IPv4 and IPv6 literals, and custom ports remain valid. A syntax error blocks both discovery and creation at the field, while a request failure remains a separate provider error. **Fetch available models** asks the `llm/discoverModels` Remote about the endpoint the form shows, so adding a provider is one pass instead of save-then-return; the reply opens a searchable picker rather than being written, and nothing is written until **Add selected**. Each selected candidate copies its id, display name, context window, output-token cap, and disclosed input types into the editable row when disclosed, while an existing row retains its user-tuned values. Search matches model ids and optional display names without clearing hidden selections. **Select all** adds the visible results, while **Deselect all** clears the entire selection so hidden results cannot be adopted accidentally. A row is deletable only when the user layer alone carries it (removal restores the composition base), and its confirmation dialog names the provider.
 
 ### First-run dialogs
 
@@ -110,7 +110,8 @@ These limits define the editor's field coverage and the page's reach; they are c
 
 - **Only the API key and curated fold fields are editable on the card** — the hand-written editor traded schema-generic field coverage for the mockup layout. Retry policy, timeouts, DeepSeek model descriptions, and other advanced fields remain in `settings.yaml`; existing model fields the editor does not show are preserved.
 - **Credential cleanup is intentionally narrow** — deleting a row removes the configured, writable credential only when its reference is the exact `<ROUTE>_API_KEY` target this page derives. Custom references, environment credentials, and unidentifiable targets are retained because the row cannot prove ownership of them.
-- **Only pi-ai routes can be hand-declared** — the custom-provider card writes into `llm-pi-ai`, the one namespace whose profiles describe a whole provider. A `llm-deepseek` route is a composition fact, not something this page can create.
+- **Only pi-ai routes can be hand-declared** — the custom-API form writes into `llm-pi-ai`, the one namespace whose profiles describe a whole provider. A `llm-deepseek` route is a composition fact, not something this page can create.
+- **The catalog select lists route identifiers** — `moonshotai`, `zai`, and the other pi-ai catalog ids appear as they are, with no product names, aliases, or search. A searchable picker with the custom form as one pinned entry would let the mode switch go; it needs the directory to carry display names first.
 - **Interrogation covers OpenAI-compatible and Anthropic Messages endpoints** — OpenAI protocols accept a standard `data` array or an enriched `models` map, while Anthropic uses its native model-listing route; every other protocol reports that it cannot be asked and its models are entered by hand.
 - **Undeclared live routes render nowhere** — a route registered without a configurable-provider declaration has no settings address; it stays visible in pickers but not on this page's rows.
 
diff --git a/packages/client/ui-settings-models/README.zh.md b/packages/client/ui-settings-models/README.zh.md
index baa0f02e82..c4ba114dbb 100644
--- a/packages/client/ui-settings-models/README.zh.md
+++ b/packages/client/ui-settings-models/README.zh.md
@@ -43,7 +43,7 @@ kind: "package-reference"
 
 ### 新增与删除提供方
 
-「新增」流程是一张承载休眠目录提供方选择框的卡片——裸挂载的 `llm-pi-ai` 在任何路由存在之前就能提供其完整的已安装 catalog。**添加自定义提供方**声明一条 pi-ai 不提供的路由；创建卡片会索要唯一的 **Provider ID**、端点、协议与至少一个可唯一识别的模型，因为没有东西能为它们兜底。端点占位示例随所选协议变化：OpenAI Chat Completions 与 Responses 显示 `https://gateway.example/v1`，Anthropic Messages 显示 `https://gateway.example`，因为其 SDK 会追加 `/v1/messages`。切换协议会保留已输入的地址。端点必须是可解析的 HTTP 或 HTTPS URL；localhost、IPv4 与 IPv6 字面地址以及自定义端口仍然有效。语法错误会在字段处阻止询问与创建，请求失败则继续作为独立的提供方错误显示。**获取可用模型**通过 `llm/discoverModels` Remote 查询表单显示的端点，因此新增提供方一次即可完成，而非先保存再返回；回复打开的是可搜索选择器而非直接写入，只有点击**添加所选**才会写入。每个选中候选会在提供方公布相应信息时，把 id、显示名、上下文窗口、最大输出 token 数和已公布的输入类型复制进可编辑行；已经存在的行保留用户调整过的值。搜索会匹配模型 id 与可选显示名称，且不会清除隐藏项的勾选状态。**全选**会加入可见结果，而**取消全选**会清空全部勾选，以免意外采用隐藏结果。只有用户层单独携带某行时，该行才可删除（删除会恢复组合基线），其确认对话框会指名该提供方。
+**添加模型提供商**是新增卡片的唯一入口：只要任一带编辑器的设置 namespace 已挂载就会出现，下述两种方式任一可继续时才可点击。卡片打开后顶部是分段式的方式切换，下方一行小字说明所选方式的用途：**第三方模型提供商**承载休眠目录提供商选择框与提供商编辑器——裸挂载的 `llm-pi-ai` 在任何路由存在之前就能提供其完整的已安装 catalog；**自定义模型 API**承载为 pi-ai 不提供的路由（中转站、自部署服务或其他兼容 OpenAI / Anthropic 协议的接口）准备的创建表单。某种方式只在其 namespace 已挂载时提供；只剩一种方式时卡片直接显示该表单；namespace 已无可采用的提供商或没有可声明的协议时，对应方式会被禁用。面板在其方式首次显示时挂载，之后在卡片打开期间保持挂载但隐藏，因此切换方式不会丢掉任何一边的草稿。自定义表单会索要唯一的 **Provider ID**、端点、协议与至少一个可唯一识别的模型，因为没有东西能为它们兜底。协议选择框以产品名称显示各协议——OpenAI Chat Completions、OpenAI Responses、Anthropic Messages——存储的值仍是 schema 标识符；适配器新增而本页尚未命名的协议直接显示其标识符。端点占位示例随所选协议变化：OpenAI Chat Completions 与 Responses 显示 `https://gateway.example/v1`，Anthropic Messages 显示 `https://gateway.example`，因为其 SDK 会追加 `/v1/messages`。切换协议会保留已输入的地址。端点必须是可解析的 HTTP 或 HTTPS URL；localhost、IPv4 与 IPv6 字面地址以及自定义端口仍然有效。语法错误会在字段处阻止询问与创建，请求失败则继续作为独立的提供方错误显示。**获取可用模型**通过 `llm/discoverModels` Remote 查询表单显示的端点，因此新增提供方一次即可完成，而非先保存再返回；回复打开的是可搜索选择器而非直接写入，只有点击**添加所选**才会写入。每个选中候选会在提供方公布相应信息时，把 id、显示名、上下文窗口、最大输出 token 数和已公布的输入类型复制进可编辑行；已经存在的行保留用户调整过的值。搜索会匹配模型 id 与可选显示名称，且不会清除隐藏项的勾选状态。**全选**会加入可见结果，而**取消全选**会清空全部勾选，以免意外采用隐藏结果。只有用户层单独携带某行时，该行才可删除（删除会恢复组合基线），其确认对话框会指名该提供方。
 
 ### 首次运行弹窗
 
@@ -110,7 +110,8 @@ kind: "package-reference"
 
 - **卡片上只有 API 密钥与精选折叠字段可编辑**：手写编辑器以 schema 通用字段覆盖换取了 mockup 布局。重试策略、超时、DeepSeek 模型说明及其他进阶字段仍留在 `settings.yaml` 中；编辑器未展示的现有模型字段会予以保留。
 - **凭据清理范围刻意保持狭窄**：删除一行时，仅当其引用与页面派生的 `<ROUTE>_API_KEY` 目标完全一致，才会清除已配置且可写的凭据。自定义引用、环境凭据与无法识别的目标会保留，因为该行无法证明自己拥有它们。
-- **只有 pi-ai 路由可以手工声明**：自定义提供方卡片写入 `llm-pi-ai`——唯一一个其 profile 描述整个提供方的 namespace。`llm-deepseek` 路由是组合面的事实，不是本页能创建的东西。
+- **只有 pi-ai 路由可以手工声明**：自定义模型 API 表单写入 `llm-pi-ai`——唯一一个其 profile 描述整个提供商的 namespace。`llm-deepseek` 路由是组合面的事实，不是本页能创建的东西。
+- **目录选择框列出的是路由标识符**：`moonshotai`、`zai` 等 pi-ai catalog id 原样显示，没有产品名、别名或搜索。一个把自定义表单作为置顶项的可搜索选择器可以取代方式切换；前提是目录先携带显示名称。
 - **询问覆盖 OpenAI 兼容与 Anthropic Messages 端点**：OpenAI 协议接受标准 `data` 数组或富信息 `models` 对象，Anthropic 则使用原生模型列表路由；其余协议会报告自己无法被询问，其模型需手工填写。
 - **未声明的存活路由无处渲染**：未附带可配置提供方声明即注册的路由没有 settings 地址；它在各选择器中仍然可见，但不会出现在本页的行里。
 
diff --git a/packages/client/ui-settings-models/src/client/CustomProviderCard.tsx b/packages/client/ui-settings-models/src/client/CustomProviderCard.tsx
index 4f382e92a1..92c1e762dd 100644
--- a/packages/client/ui-settings-models/src/client/CustomProviderCard.tsx
+++ b/packages/client/ui-settings-models/src/client/CustomProviderCard.tsx
@@ -3,12 +3,13 @@
  * gateway, a self-hosted server, or a provider newer than the installed
  * catalog.
  *
- * This is a create, not an edit, which is why it is its own card rather than
+ * This is a create, not an edit, which is why it is its own form rather than
  * the provider editor with extra fields: the route id is being *chosen* here,
- * and the settings address does not exist until it is. One `settings.mutate`
- * sets the whole profile at `providers.<route>`; the key travels separately
- * through `credentials/set` under the reference the profile records, exactly as
- * an existing provider's key does.
+ * and the settings address does not exist until it is. It renders as the
+ * custom-API panel of the section's add card, whose mode switch is its title.
+ * One `settings.mutate` sets the whole profile at `providers.<route>`; the key
+ * travels separately through `credentials/set` under the reference the profile
+ * records, exactly as an existing provider's key does.
  *
  * The three fields a hand-declared route cannot default — endpoint, protocol,
  * and at least one model — are required here rather than at load, so the
@@ -30,6 +31,7 @@ import { validateDeepSeekModels } from './DeepSeekModelsEditor.tsx'
 import { ModelListEditor } from './ModelListEditor.tsx'
 import type { ModelDraft } from './ModelListEditor.tsx'
 import { deriveKeyRef } from './store.ts'
+import { protocolLabel } from './protocol-label.ts'
 import type { ModelsOperations } from './operations.ts'
 import type { en } from './locales.ts'
 import styles from './ModelsSection.module.css'
@@ -198,9 +200,6 @@ export function CustomProviderCard(props: CustomProviderCardProps): ReactNode {
 
   return (
     <div className={styles['editor']}>
-      <div className={styles['editorHeader']}>
-        <span className={styles['editorTitle']}>{t('customTitle')}</span>
-      </div>
       <div className={styles['field']}>
         <span className={styles['fieldLabel']}>{t('customRoute')}</span>
         <input
@@ -255,7 +254,7 @@ export function CustomProviderCard(props: CustomProviderCardProps): ReactNode {
           disabled={profileDisabled}
           onChange={(event) => { setProtocol(event.target.value) }}
         >
-          {protocols.map(choice => <option key={choice} value={choice}>{choice}</option>)}
+          {protocols.map(choice => <option key={choice} value={choice}>{protocolLabel(t, choice)}</option>)}
         </select>
       </div>
       <div className={styles['field']}>
diff --git a/packages/client/ui-settings-models/src/client/ModelsSection.module.css b/packages/client/ui-settings-models/src/client/ModelsSection.module.css
index 15b8933fa3..d39e807876 100644
--- a/packages/client/ui-settings-models/src/client/ModelsSection.module.css
+++ b/packages/client/ui-settings-models/src/client/ModelsSection.module.css
@@ -305,20 +305,16 @@
   gap: 12px;
 }
 
-/* The two ways to gain a provider, as equal siblings spanning the same width
-   as the rows above. Wraps rather than shrinking below a legible label. */
+/* The one entry to the add card, spanning the same width as the rows above. */
 .addActions {
   display: flex;
-  flex-wrap: wrap;
-  gap: 10px;
 }
 
 .addButton {
-  /* Overrides the shared button base above: these two are not pills sitting in
-     a footer but the last slot of the provider list, so they split the row
-     evenly and repeat the row cards' corner. Dashed, like every other "nothing
-     here yet" affordance on this page, to read as a place rather than a
-     command. */
+  /* Overrides the shared button base above: this is not a pill sitting in a
+     footer but the last slot of the provider list, so it fills the row and
+     repeats the row cards' corner. Dashed, like every other "nothing here yet"
+     affordance on this page, to read as a place rather than a command. */
   flex: 1 1 0;
   min-width: 180px;
   gap: 6px;
@@ -327,6 +323,28 @@
   border-radius: 16px;
 }
 
+/* The add card's mode switch with the one-line purpose of the chosen mode
+   beneath it; the switch hugs its labels rather than spanning the card. */
+.addModes {
+  display: flex;
+  flex-direction: column;
+  align-items: flex-start;
+  gap: 8px;
+}
+
+/* One panel per mode. Both stay mounted while the card is open; the hidden
+   attribute must win over the panel's own display, or the inactive draft
+   would show beneath the active one. */
+.addPanel {
+  display: flex;
+  flex-direction: column;
+  gap: 14px;
+}
+
+.addPanel[hidden] {
+  display: none;
+}
+
 .addCard,
 .setupCard {
   border-radius: 12px;
diff --git a/packages/client/ui-settings-models/src/client/ModelsSection.tsx b/packages/client/ui-settings-models/src/client/ModelsSection.tsx
index a9ce163a50..ce39f9f8cc 100644
--- a/packages/client/ui-settings-models/src/client/ModelsSection.tsx
+++ b/packages/client/ui-settings-models/src/client/ModelsSection.tsx
@@ -5,16 +5,21 @@
  * solid configured or missing dots. A whole-section provider without a
  * configured key renders as its open setup card instead of a row, but only in
  * the first-run posture — no provider on the page can serve requests yet — and
- * only until the user closes that card; the add flow is a card carrying the
- * dormant-provider select. Each card kind owns its own open state, so closing
- * one never discards a draft in another. Every mutation writes through the
- * wire, while a provider removal first requires confirmation; the page
- * re-renders from pushed invalidations or the post-apply reload.
+ * only until the user closes that card. The add flow is one card behind one
+ * button: a mode switch chooses between adopting a dormant directory provider
+ * (the catalog select over the provider editor) and declaring a custom model
+ * API (the create form). A panel mounts the first time its mode is shown and
+ * stays mounted, hidden, while the card is open, so switching modes discards
+ * neither draft and an unvisited mode costs nothing. Each card kind owns its
+ * own open state, so closing one never discards a draft in another. Every
+ * mutation writes through the wire, while a provider removal first requires
+ * confirmation; the page re-renders from pushed invalidations or the
+ * post-apply reload.
  */
 
 import { useState } from 'react'
 import type { ReactNode } from 'react'
-import { Button, IconPlusOutlineRegular, Modal } from '@deepseek-ai/dsh-client-ui-primitives'
+import { Button, IconPlusOutlineRegular, Modal, SegmentedControl } from '@deepseek-ai/dsh-client-ui-primitives'
 import type { InjectFace, PropsRenderSlots } from '@deepseek-ai/dsh-client-ui-slots'
 // Type-only: pulls this package's SlotMap merge (the two Models child slots).
 import type {} from './slot-contract.ts'
@@ -43,6 +48,12 @@ export interface ModelsSectionInjected {
   t: (key: keyof typeof en) => string
 }
 
+/**
+ * The two ways the add card gains a provider: adopt a directory row the
+ * adapter already knows, or declare a route it does not.
+ */
+type AddMode = 'catalog' | 'custom'
+
 /** The child slots this section declares and dispatches (see ./slot-contract.ts). */
 type ModelsChildSlots = 'settings.models.provider-card' | 'settings.models.footer'
 
@@ -205,12 +216,14 @@ function Loaded({ injected, renderSlot }: { injected: ModelsSectionFace; renderS
   const { controller, operations, schema, t } = injected
   const state = injected.useSnapshot(snapshot => snapshot)
   const [editing, setEditing] = useState<EditorTarget | undefined>(undefined)
-  const [adding, setAdding] = useState(false)
+  const [addOpen, setAddOpen] = useState(false)
+  const [addMode, setAddMode] = useState<AddMode>('catalog')
+  /** The modes shown since the add card opened; each keeps its panel mounted. */
+  const [visited, setVisited] = useState<ReadonlySet<AddMode>>(() => new Set())
   const [deleteTarget, setDeleteTarget] = useState<EditorTarget | undefined>(undefined)
   const [deleting, setDeleting] = useState(false)
   const [deleteFailure, setDeleteFailure] = useState<string | undefined>(undefined)
   const [savedTarget, setSavedTarget] = useState<ProviderIdentity | undefined>(undefined)
-  const [declaring, setDeclaring] = useState(false)
   const [dismissedSetup, setDismissedSetup] = useState<ReadonlySet<string>>(() => new Set())
 
   const announceSaved = (target: ProviderIdentity): void => {
@@ -222,17 +235,16 @@ function Loaded({ injected, renderSlot }: { injected: ModelsSectionFace; renderS
 
   const closeEditor = (changed: boolean, target: ProviderIdentity): void => {
     setEditing(undefined)
-    setAdding(false)
-    setDeclaring(false)
+    setAddOpen(false)
     if (changed) announceSaved(target)
   }
 
   /**
-   * Close a setup card, which owns none of the state above: the row-editor,
-   * add, and declare cards each own one of those, so clearing them here would
-   * discard a draft the user opened beside this card. Dismissal is this card's
-   * own — the provider falls back to an ordinary row for the rest of the
-   * session, and reopens through Edit.
+   * Close a setup card, which owns none of the state above: the row-editor
+   * and add cards each own one of those, so clearing them here would discard
+   * a draft the user opened beside this card. Dismissal is this card's own —
+   * the provider falls back to an ordinary row for the rest of the session,
+   * and reopens through Edit.
    */
   const closeSetup = (changed: boolean, target: ProviderIdentity): void => {
     setDismissedSetup(previous => new Set([...previous, target.provider]))
@@ -292,7 +304,19 @@ function Loaded({ injected, renderSlot }: { injected: ModelsSectionFace; renderS
   const configured = state.rows.filter(row => row.configured)
   const configurable = state.rows.filter(row => state.namespaces.has(row.entry.settingsNs))
   const addable = configurable.filter(row => !row.configured)
-  const addTarget = adding ? editing : undefined
+  // Hand-declared routes live in the pi-ai namespace, which is also the only
+  // one whose schema names the protocols one may speak; without it mounted
+  // there is nothing to declare and the mode is not offered.
+  const protocols = protocolChoices(state.namespaces.get('llm-pi-ai'), schema)
+  // Each mode is offered while its namespace is mounted and enabled while it
+  // has something to offer; the card shows the chosen mode where both are
+  // offered, else the only one there is.
+  const catalogOffered = configurable.length > 0
+  const catalogEnabled = addable.length > 0
+  const customOffered = state.namespaces.has('llm-pi-ai')
+  const customEnabled = protocols.length > 0
+  const mode: AddMode = catalogOffered && customOffered ? addMode : customOffered ? 'custom' : 'catalog'
+  const addTarget = addOpen && catalogOffered ? editing : undefined
   const addNamespace = addTarget === undefined ? undefined : state.namespaces.get(addTarget.settingsNs)
   // The draft's directory row, for the card extension seat. A refresh can drop
   // the row mid-draft (the route was adopted or withdrawn elsewhere); the
@@ -300,10 +324,6 @@ function Loaded({ injected, renderSlot }: { injected: ModelsSectionFace; renderS
   const addRow = addTarget === undefined
     ? undefined
     : state.rows.find(row => row.entry.provider === addTarget.provider)
-  // Hand-declared routes live in the pi-ai namespace, which is also the only
-  // one whose schema names the protocols one may speak; without it mounted
-  // there is nothing to declare and the entry point stays disabled.
-  const protocols = protocolChoices(state.namespaces.get('llm-pi-ai'), schema)
 
   return (
     <div className={styles['section']}>
@@ -349,7 +369,7 @@ function Loaded({ injected, renderSlot }: { injected: ModelsSectionFace; renderS
               </li>
             )
           }
-          const open = !adding && editing?.provider === row.entry.provider
+          const open = !addOpen && editing?.provider === row.entry.provider
           const credentialConfigured = row.credential?.configured === true
           const credentialMissing = !credentialConfigured
             && row.apiKeyEnv !== undefined
@@ -392,11 +412,10 @@ function Loaded({ injected, renderSlot }: { injected: ModelsSectionFace; renderS
                     aria-label={providerCopy(t('editProvider'), target)}
                     onClick={() => {
                       setSavedTarget(undefined)
-                      // One card at a time: leaving `declaring` set would show
-                      // the create card beside this editor, and closing either
-                      // one discards the other's draft.
-                      setDeclaring(false)
-                      setAdding(false)
+                      // One card at a time: the add card closes with whatever
+                      // it held, since closing either card would otherwise
+                      // discard the other's draft.
+                      setAddOpen(false)
                       setEditing(open ? undefined : target)
                     }}
                   >
@@ -443,110 +462,130 @@ function Loaded({ injected, renderSlot }: { injected: ModelsSectionFace; renderS
         })}
       </ul>
       <div className={styles['addBlock']}>
-        {addTarget !== undefined && addNamespace !== undefined
+        {addOpen
           ? (
             <div className={styles['addCard']}>
-              <div className={styles['field']}>
-                <span className={styles['fieldLabel']}>{t('provider')}</span>
-                <select
-                  className={`${styles['input']} ${styles['selectInput']}`}
-                  value={addTarget.provider}
-                  aria-label={t('provider')}
-                  onChange={(event) => {
-                    const row = addable.find(candidate => candidate.entry.provider === event.target.value)
-                    /* v8 ignore next -- the select only lists addable rows */
-                    if (row === undefined) return
-                    setEditing(targetOf(row))
-                  }}
-                >
-                  {addable.map(row => (
-                    <option key={row.entry.provider} value={row.entry.provider}>{row.entry.displayName}</option>
-                  ))}
-                </select>
-              </div>
-              <ProviderEditor
-                key={addTarget.provider}
-                provider={addTarget.provider}
-                displayName={addTarget.displayName}
-                hideTitle
-                namespace={addNamespace}
-                schema={schema}
-                settingsPath={addTarget.settingsPath}
-                operations={operations}
-                t={t}
-                readOnly={!state.writable}
-                onClose={(changed) => { closeEditor(changed, addTarget) }}
-              />
-              {addRow === undefined
-                ? null
-                : renderSlot(
-                  'settings.models.provider-card',
-                  { provider: addRow.entry, configured: addRow.configured, keyConfigured: keyConfiguredOf(addRow) },
-                  { entryKey: addRow.entry.settingsNs },
-                )}
+              {catalogOffered && customOffered
+                ? (
+                  <div className={styles['addModes']}>
+                    <SegmentedControl
+                      label={t('addMode')}
+                      value={mode}
+                      options={[
+                        { value: 'catalog', label: t('addCatalog'), disabled: !catalogEnabled },
+                        { value: 'custom', label: t('addCustom'), disabled: !customEnabled },
+                      ]}
+                      onChange={(next) => {
+                        setAddMode(next)
+                        setVisited(previous => new Set([...previous, next]))
+                      }}
+                    />
+                    <p className={styles['advancedHint']}>
+                      {t(mode === 'catalog' ? 'addCatalogHint' : 'addCustomHint')}
+                    </p>
+                  </div>
+                )
+                : null}
+              {visited.has('catalog') && addTarget !== undefined && addNamespace !== undefined
+                ? (
+                  <div
+                    role="tabpanel"
+                    aria-label={t('addCatalog')}
+                    hidden={mode !== 'catalog'}
+                    className={styles['addPanel']}
+                  >
+                    <div className={styles['field']}>
+                      <span className={styles['fieldLabel']}>{t('provider')}</span>
+                      <select
+                        className={`${styles['input']} ${styles['selectInput']}`}
+                        value={addTarget.provider}
+                        aria-label={t('provider')}
+                        onChange={(event) => {
+                          const row = addable.find(candidate => candidate.entry.provider === event.target.value)
+                          /* v8 ignore next -- the select only lists addable rows */
+                          if (row === undefined) return
+                          setEditing(targetOf(row))
+                        }}
+                      >
+                        {addable.map(row => (
+                          <option key={row.entry.provider} value={row.entry.provider}>{row.entry.displayName}</option>
+                        ))}
+                      </select>
+                    </div>
+                    <ProviderEditor
+                      key={addTarget.provider}
+                      provider={addTarget.provider}
+                      displayName={addTarget.displayName}
+                      hideTitle
+                      namespace={addNamespace}
+                      schema={schema}
+                      settingsPath={addTarget.settingsPath}
+                      operations={operations}
+                      t={t}
+                      readOnly={!state.writable}
+                      onClose={(changed) => { closeEditor(changed, addTarget) }}
+                    />
+                    {addRow === undefined
+                      ? null
+                      : renderSlot(
+                        'settings.models.provider-card',
+                        { provider: addRow.entry, configured: addRow.configured, keyConfigured: keyConfiguredOf(addRow) },
+                        { entryKey: addRow.entry.settingsNs },
+                      )}
+                  </div>
+                )
+                : null}
+              {visited.has('custom')
+                ? (
+                  <div
+                    role="tabpanel"
+                    aria-label={t('addCustom')}
+                    hidden={mode !== 'custom'}
+                    className={styles['addPanel']}
+                  >
+                    <CustomProviderCard
+                      taken={state.rows.map(row => row.entry.provider)}
+                      protocols={protocols}
+                      /* v8 ignore next -- the panel only renders with this namespace mounted */
+                      revision={state.namespaces.get('llm-pi-ai')?.revision ?? 0}
+                      operations={operations}
+                      t={t}
+                      readOnly={!state.writable}
+                      onClose={(changed) => {
+                        setAddOpen(false)
+                        if (changed) void controller.load()
+                      }}
+                    />
+                  </div>
+                )
+                : null}
             </div>
           )
-          : declaring
+          : catalogOffered || customOffered
             ? (
-              <div className={styles['addCard']}>
-                <CustomProviderCard
-                  taken={state.rows.map(row => row.entry.provider)}
-                  protocols={protocols}
-                  /* v8 ignore next -- the card only opens from a button disabled without this namespace */
-                  revision={state.namespaces.get('llm-pi-ai')?.revision ?? 0}
-                  operations={operations}
-                  t={t}
-                  readOnly={!state.writable}
-                  onClose={(changed) => {
-                    setDeclaring(false)
-                    if (changed) void controller.load()
+              // One entry for both ways to gain a provider; the card behind it
+              // splits them. Full width, so it lines up with the rows above.
+              <div className={styles['addActions']}>
+                <button
+                  type="button"
+                  className={styles['addButton']}
+                  disabled={!state.writable || (!catalogEnabled && !customEnabled)}
+                  onClick={() => {
+                    const first = addable[0]
+                    const initial: AddMode = catalogEnabled ? 'catalog' : 'custom'
+                    setSavedTarget(undefined)
+                    setEditing(first === undefined ? undefined : targetOf(first))
+                    setAddMode(initial)
+                    setVisited(new Set([initial]))
+                    setAddOpen(true)
                   }}
-                />
+                >
+                  <IconPlusOutlineRegular size={14} />
+                  {t('add')}
+                </button>
               </div>
             )
-            : (
-              // One row for the two ways to gain a provider: adopt one the
-              // adapter already knows, or declare one it does not. Side by side
-              // and equal-width so they read as siblings and line up with the
-              // rows above, rather than two pills of different lengths.
-              <div className={styles['addActions']}>
-                {configurable.length > 0 && (
-                  <button
-                    type="button"
-                    className={styles['addButton']}
-                    disabled={addable.length === 0 || !state.writable}
-                    onClick={() => {
-                      const first = addable[0]
-                      /* v8 ignore next -- the button is disabled while nothing is addable */
-                      if (first === undefined) return
-                      setSavedTarget(undefined)
-                      setDeclaring(false)
-                      setAdding(true)
-                      setEditing(targetOf(first))
-                    }}
-                  >
-                    <IconPlusOutlineRegular size={14} />
-                    {t('add')}
-                  </button>
-                )}
-                {state.namespaces.has('llm-pi-ai') && (
-                  <button
-                    type="button"
-                    className={styles['addButton']}
-                    disabled={protocols.length === 0 || !state.writable}
-                    onClick={() => {
-                      setSavedTarget(undefined)
-                      setAdding(false)
-                      setEditing(undefined)
-                      setDeclaring(true)
-                    }}
-                  >
-                    <IconPlusOutlineRegular size={14} />
-                    {t('customAdd')}
-                  </button>
-                )}
-              </div>
-            )}
+            : null}
       </div>
       {renderSlot('settings.models.footer', {})}
       <Modal
diff --git a/packages/client/ui-settings-models/src/client/ProviderEditor.tsx b/packages/client/ui-settings-models/src/client/ProviderEditor.tsx
index bc89f43fc8..4efd1447fa 100644
--- a/packages/client/ui-settings-models/src/client/ProviderEditor.tsx
+++ b/packages/client/ui-settings-models/src/client/ProviderEditor.tsx
@@ -34,6 +34,7 @@ import { apiKeyFailure } from './apiKey.ts'
 import { EditorFooter } from './EditorFooter.tsx'
 import { ModelListEditor } from './ModelListEditor.tsx'
 import { deriveKeyRef, protocolChoices } from './store.ts'
+import { protocolLabel } from './protocol-label.ts'
 import type { ModelsOperations } from './operations.ts'
 import type { SettingsSchemaOperations } from './schema-operations.ts'
 import type { en } from './locales.ts'
@@ -445,7 +446,7 @@ export function ProviderEditor(props: ProviderEditorProps): ReactNode {
                         reader announces it either way, and an empty one is
                         announced as a choice with no identity. */}
                     {probeApi === undefined ? <option value="">{t('customApiUnset')}</option> : null}
-                    {protocols.map(choice => <option key={choice} value={choice}>{choice}</option>)}
+                    {protocols.map(choice => <option key={choice} value={choice}>{protocolLabel(t, choice)}</option>)}
                   </select>
                 </div>
               )
diff --git a/packages/client/ui-settings-models/src/client/locales.ts b/packages/client/ui-settings-models/src/client/locales.ts
index 2b0f92d8df..9cd6d52321 100644
--- a/packages/client/ui-settings-models/src/client/locales.ts
+++ b/packages/client/ui-settings-models/src/client/locales.ts
@@ -14,7 +14,12 @@ export const en = {
   deleteDescriptionWithCredential: 'Deleting {provider} removes its configuration and stored API key.',
   deleteConfirm: 'Delete {provider}',
   deleting: 'Deleting {provider}…',
-  add: 'Add provider',
+  add: 'Add model provider',
+  addMode: 'How to add',
+  addCatalog: 'Third-party model provider',
+  addCustom: 'Custom model API',
+  addCatalogHint: 'Pick OpenAI, Anthropic, Kimi, or another provider from the built-in catalog and enter its API key.',
+  addCustomHint: 'Connect a relay, a self-hosted server, or any other OpenAI- or Anthropic-compatible endpoint by its base URL, protocol, and models.',
   provider: 'Provider',
   close: 'Close',
   cancel: 'Cancel',
@@ -79,8 +84,6 @@ export const en = {
   fetchSelectAll: 'Select all',
   fetchDeselectAll: 'Deselect all',
   fetchAdopt: 'Add selected',
-  customAdd: 'Add a custom provider',
-  customTitle: 'Custom provider',
   customTag: 'Custom',
   customRoute: 'Provider ID',
   customRouteHint: 'Lowercase identifier, starting with a letter, that uniquely names this provider in requests and as its credential name.',
@@ -89,6 +92,9 @@ export const en = {
   customDisplayName: 'Display name',
   customApi: 'API protocol',
   customApiUnset: 'Not selected',
+  protocolOpenAiCompletions: 'OpenAI Chat Completions',
+  protocolOpenAiResponses: 'OpenAI Responses',
+  protocolAnthropicMessages: 'Anthropic Messages',
   customNeedsBaseUrl: 'A custom provider needs a base URL.',
   customBaseUrlInvalid: 'Enter a valid HTTP or HTTPS URL.',
   customNeedsModels: 'A custom provider needs at least one model.',
@@ -116,7 +122,7 @@ export type ModelsKey = keyof typeof en
 export const zh: { [Key in keyof typeof en]: string } = {
   nav: '模型',
   title: '模型',
-  intro: '填入各提供方的 API 密钥即可使用其模型。',
+  intro: '填入各提供商的 API 密钥即可使用其模型。',
   edit: '编辑',
   editProvider: '编辑 {provider}',
   remove: '删除',
@@ -126,8 +132,13 @@ export const zh: { [Key in keyof typeof en]: string } = {
   deleteDescriptionWithCredential: '删除 {provider} 会移除其配置和存储的 API 密钥。',
   deleteConfirm: '删除 {provider}',
   deleting: '正在删除 {provider}…',
-  add: '添加提供方',
-  provider: '提供方',
+  add: '添加模型提供商',
+  addMode: '添加方式',
+  addCatalog: '第三方模型提供商',
+  addCustom: '自定义模型 API',
+  addCatalogHint: '从内置目录中选择 OpenAI、Anthropic、Kimi 等提供商，填入其 API 密钥即可使用。',
+  addCustomHint: '连接中转站、自部署服务或其他兼容 OpenAI / Anthropic 协议的接口，需填写 API 地址、协议和模型。',
+  provider: '提供商',
   close: '关闭',
   cancel: '取消',
   apply: '保存',
@@ -136,7 +147,7 @@ export const zh: { [Key in keyof typeof en]: string } = {
   credentialConfigured: 'API 密钥已配置',
   credentialMissing: 'API 密钥缺失',
   readOnly: '当前部署的设置文档为只读。',
-  loadFailed: '加载提供方目录失败',
+  loadFailed: '加载提供商目录失败',
   conflict: '这张卡片打开期间，这些设置已被其他地方改动。请关闭后重新打开，在当前值上编辑。',
   retry: '重试',
   keyInput: 'API 密钥',
@@ -146,7 +157,7 @@ export const zh: { [Key in keyof typeof en]: string } = {
   keyEnvLocked: '由启动环境提供（只读）',
   customized: '自定义设置',
   baseUrl: 'API 地址',
-  baseUrlDefault: '提供方默认',
+  baseUrlDefault: '提供商默认',
   deepSeekChatBaseUrl: 'https://api.deepseek.com',
   deepSeekMessagesBaseUrl: 'https://api.deepseek.com/anthropic',
   deepSeekEndpointHint: '请填写与当前连接配置兼容的 API 地址。',
@@ -159,9 +170,9 @@ export const zh: { [Key in keyof typeof en]: string } = {
   modelName: '显示名称',
   modelNamePlaceholder: '留空时使用模型 ID',
   contextWindow: '上下文窗口',
-  contextWindowPlaceholder: '使用提供方默认值',
+  contextWindowPlaceholder: '使用提供商默认值',
   maxTokens: '最大输出 token 数',
-  maxTokensPlaceholder: '使用提供方默认值',
+  maxTokensPlaceholder: '使用提供商默认值',
   modelAdvanced: '模型选项',
   modelInputTypes: '输入类型',
   modelInputText: '文本',
@@ -170,7 +181,7 @@ export const zh: { [Key in keyof typeof en]: string } = {
   removeModel: '删除模型',
   modelsEmpty: '模型选择器中将不显示任何模型；目录外 ID 仍可直接发送。',
   keyBlank: '请输入 API 密钥；留空则保持已存储的密钥。',
-  keyBlankNew: '请输入 API 密钥；若该提供方以其他方式鉴权，可以留空。',
+  keyBlankNew: '请输入 API 密钥；若该提供商以其他方式鉴权，可以留空。',
   keyIllegalCharacters: '该 API 密钥格式错误，请检查。',
   modelIdRequired: '模型 ID 不能为空。',
   modelIdDuplicate: '模型 ID 不能重复。',
@@ -181,33 +192,34 @@ export const zh: { [Key in keyof typeof en]: string } = {
   modelCapacityInvalid: '容量需为数字，可加 K 或 M 后缀。',
   modelDuplicate: '每个模型 ID 只能出现一次。',
   fetchModels: '获取可用模型',
-  fetching: '正在询问提供方\u2026',
+  fetching: '正在询问提供商\u2026',
   fetchNeedsBaseUrl: '请先填写 API 地址，再获取。',
-  fetchEmpty: '该提供方没有列出任何模型，请手动添加。',
+  fetchEmpty: '该提供商没有列出任何模型，请手动添加。',
   fetchTitle: '选择要添加的模型',
-  fetchDescription: '以下是模型提供方的可用模型，勾选要添加的模型。',
+  fetchDescription: '以下是模型提供商的可用模型，勾选要添加的模型。',
   fetchSearch: '搜索模型',
   fetchNoMatches: '没有匹配的模型。',
   fetchSelectAll: '全选',
   fetchDeselectAll: '取消全选',
   fetchAdopt: '添加所选',
-  customAdd: '添加自定义提供方',
-  customTitle: '自定义提供方',
   customTag: '自定义',
   customRoute: 'Provider ID',
-  customRouteHint: '以小写字母开头的标识，在请求中唯一标识该提供方，并用于派生凭据名。',
+  customRouteHint: '以小写字母开头的标识，在请求中唯一标识该提供商，并用于派生凭据名。',
   customRouteInvalid: '需以小写字母开头，之后可用小写字母、数字和短横线。',
-  customRouteTaken: '已有提供方使用了这个 ID。',
+  customRouteTaken: '已有提供商使用了这个 ID。',
   customDisplayName: '显示名称',
   customApi: 'API 协议',
   customApiUnset: '未选择',
-  customNeedsBaseUrl: '自定义提供方需要填写 API 地址。',
+  protocolOpenAiCompletions: 'OpenAI Chat Completions',
+  protocolOpenAiResponses: 'OpenAI Responses',
+  protocolAnthropicMessages: 'Anthropic Messages',
+  customNeedsBaseUrl: '自定义模型 API 需要填写 API 地址。',
   customBaseUrlInvalid: '请输入有效的 HTTP 或 HTTPS 地址。',
-  customNeedsModels: '自定义提供方至少需要一个模型。',
+  customNeedsModels: '自定义模型 API 至少需要一个模型。',
   customBaseUrlPlaceholder: 'https://gateway.example/v1',
   customAnthropicBaseUrlPlaceholder: 'https://gateway.example',
   settingsPathUnresolvable: '无法解析设置路径',
-  create: '创建提供方',
+  create: '创建提供商',
   creating: '创建中\u2026',
   welcomeTitle: '内测声明',
   welcomeBody: 'DeepSeek Harness 目前的 0.1 版本仍处在面向 Harness 开发者进行测试的阶段，还有许多地方需要持续改进和打磨，希望听取广大开发者的反馈建议。预计 DeepSeek Harness 的核心插件以及基础 API 都会在接下来的一段时间内快速迭代、持续演化。\n\n我们期待与全球开发者一起，在开源、开放、可复用、可组合的基础设施之上，共同探索智能上限。欢迎全球 Harness 开发者加入 DSH 插件生态。',
diff --git a/packages/client/ui-settings-models/src/client/protocol-label.ts b/packages/client/ui-settings-models/src/client/protocol-label.ts
new file mode 100644
index 0000000000..f65323f32c
--- /dev/null
+++ b/packages/client/ui-settings-models/src/client/protocol-label.ts
@@ -0,0 +1,27 @@
+/**
+ * Product names for the wire protocols a pi-ai route may speak. The pickers
+ * show these instead of the schema identifiers (`openai-completions`), which
+ * are what `settings.yaml` carries and what the option values stay.
+ */
+
+import type { ModelsKey } from './locales.ts'
+
+/** The protocols this page names, keyed by their schema identifier. */
+const PROTOCOL_LABEL_KEYS: Readonly<Record<string, ModelsKey>> = {
+  'openai-completions': 'protocolOpenAiCompletions',
+  'openai-responses': 'protocolOpenAiResponses',
+  'anthropic-messages': 'protocolAnthropicMessages',
+}
+
+/**
+ * The picker text for one protocol identifier.
+ * @param t - section copy.
+ * @param protocol - the schema identifier of the protocol.
+ * @returns the product name this page knows the protocol by; a protocol the
+ * adapter adds before this page names it shows its identifier, the spelling
+ * `settings.yaml` needs anyway.
+ */
+export function protocolLabel(t: (key: ModelsKey) => string, protocol: string): string {
+  const key = PROTOCOL_LABEL_KEYS[protocol]
+  return key === undefined ? protocol : t(key)
+}
diff --git a/packages/client/ui-settings-models/tests/components.client.spec.tsx b/packages/client/ui-settings-models/tests/components.client.spec.tsx
index e9cec03e60..125209e81b 100644
--- a/packages/client/ui-settings-models/tests/components.client.spec.tsx
+++ b/packages/client/ui-settings-models/tests/components.client.spec.tsx
@@ -46,6 +46,7 @@ function capacityInputs(label: string): HTMLInputElement[] {
 const PiAiConfig = Schema.object({
   providers: Schema.dict(Schema.object({
     apiKeyEnv: Schema.string().role('credential-ref'),
+    api: Schema.union(['openai-completions', 'openai-responses', 'anthropic-messages']),
     baseURL: Schema.string(),
     reasoning: Schema.union(['off', 'minimal', 'low', 'medium', 'high', 'xhigh', 'max']),
     headers: Schema.dict(Schema.string()),
@@ -307,12 +308,11 @@ async function mountDeepSeekCard(overrides: Parameters<typeof scriptedFace>[0] =
 }
 
 describe('ModelsSection', () => {
-  it('hides both add actions when their settings namespaces are absent', async () => {
+  it('hides the add action when no settings namespace can open an editor', async () => {
     const scripted = scriptedFace()
     scripted.face.settings.describe.mockResolvedValue(remoteOk({ writable: true, hasDocument: false, namespaces: [] }))
     await mountFace(scripted)
     expect(screen.queryByRole('button', { name: en.add })).toBeNull()
-    expect(screen.queryByRole('button', { name: en.customAdd })).toBeNull()
   })
 
   it('offers only providers whose settings namespace can open an editor', async () => {
@@ -322,8 +322,11 @@ describe('ModelsSection', () => {
       namespaces: wireNamespaces().filter(view => view.ns !== 'llm-pi-ai'),
     }))
     await mountFace(scripted)
-    expect(screen.queryByRole('button', { name: en.customAdd })).toBeNull()
     fireEvent.click(screen.getByRole('button', { name: en.add }))
+    // Without the pi-ai namespace nothing can be hand-declared, so the card
+    // is the catalog form alone: no mode switch, no custom panel.
+    expect(screen.queryByRole('tablist')).toBeNull()
+    expect(screen.queryByRole('textbox', { name: en.customRoute })).toBeNull()
     expect(screen.queryByRole('option', { name: 'anthropic' })).toBeNull()
     expect(screen.getByRole('option', { name: 'plain' })).toBeTruthy()
   })
@@ -339,7 +342,6 @@ describe('ModelsSection', () => {
     fireEvent.click(screen.getByRole('button', { name: openaiCopy(en.editProvider) }))
     expect(await screen.findByLabelText(en.keyInput)).toBeTruthy()
     expect(screen.getByRole('button', { name: en.add })).toBeTruthy()
-    expect(screen.getByRole('button', { name: en.customAdd })).toBeTruthy()
   })
 
   it('renders nothing before the slot injects its dependencies', () => {
@@ -1437,6 +1439,88 @@ describe('ModelsSection', () => {
     expect(mutate).not.toHaveBeenCalled()
   })
 
+  it('opens the add card on the catalog mode with both modes offered', async () => {
+    await mountSection()
+    fireEvent.click(screen.getByRole('button', { name: en.add }))
+    const modes = screen.getByRole('tablist', { name: en.addMode })
+    expect(within(modes).getByRole('tab', { name: en.addCatalog }).getAttribute('aria-selected')).toBe('true')
+    expect(within(modes).getByRole('tab', { name: en.addCustom }).getAttribute('aria-selected')).toBe('false')
+    expect(screen.getByText(en.addCatalogHint)).toBeTruthy()
+    expect(screen.queryByText(en.addCustomHint)).toBeNull()
+    expect(screen.getByRole('combobox', { name: en.provider })).toBeTruthy()
+    // The custom panel is mounted but hidden, so its fields are not reachable.
+    expect(screen.queryByRole('textbox', { name: en.customRoute })).toBeNull()
+  })
+
+  it('switches between the modes without losing either draft', async () => {
+    await mountSection()
+    fireEvent.click(screen.getByRole('button', { name: en.add }))
+    fireEvent.change(screen.getByLabelText(en.keyInput), { target: { value: 'sk-catalog' } })
+
+    fireEvent.click(screen.getByRole('tab', { name: en.addCustom }))
+    expect(screen.getByText(en.addCustomHint)).toBeTruthy()
+    expect(screen.queryByText(en.addCatalogHint)).toBeNull()
+    expect(screen.queryByRole('combobox', { name: en.provider })).toBeNull()
+    fireEvent.change(screen.getByRole('textbox', { name: en.customRoute }), { target: { value: 'acme' } })
+
+    // Both panels stay mounted from here on, so the shown one is queried by name.
+    fireEvent.click(screen.getByRole('tab', { name: en.addCatalog }))
+    const catalog = screen.getByRole('tabpanel', { name: en.addCatalog })
+    expect(within(catalog).getByLabelText<HTMLInputElement>(en.keyInput).value).toBe('sk-catalog')
+    fireEvent.click(screen.getByRole('tab', { name: en.addCustom }))
+    expect(screen.getByRole<HTMLInputElement>('textbox', { name: en.customRoute }).value).toBe('acme')
+  })
+
+  it('opens on the custom mode when every catalog provider is already configured', async () => {
+    const scripted = scriptedFace()
+    scripted.face.llm.listConfigurableProviders.mockResolvedValue(remoteOk([
+      { provider: 'deepseek-official', displayName: 'DeepSeek', settingsNs: 'llm-deepseek', settingsPath: [] },
+      { provider: 'openai', displayName: 'openai', settingsNs: 'llm-pi-ai', settingsPath: ['providers', 'openai'] },
+    ]))
+    await mountFace(scripted)
+    fireEvent.click(screen.getByRole('button', { name: en.add }))
+    const catalog = screen.getByRole<HTMLButtonElement>('tab', { name: en.addCatalog })
+    expect(catalog.disabled).toBe(true)
+    expect(screen.getByRole('tab', { name: en.addCustom }).getAttribute('aria-selected')).toBe('true')
+    expect(screen.getByRole('textbox', { name: en.customRoute })).toBeTruthy()
+  })
+
+  it('shows the custom form alone when no directory row can be adopted', async () => {
+    const scripted = scriptedFace()
+    scripted.face.llm.listConfigurableProviders.mockResolvedValue(remoteOk([]))
+    await mountFace(scripted)
+    fireEvent.click(screen.getByRole('button', { name: en.add }))
+    expect(screen.queryByRole('tablist')).toBeNull()
+    expect(screen.queryByRole('combobox', { name: en.provider })).toBeNull()
+    expect(screen.getByRole('textbox', { name: en.customRoute })).toBeTruthy()
+  })
+
+  it('disables the add action when neither mode can proceed', async () => {
+    const scripted = scriptedFace()
+    scripted.face.llm.listConfigurableProviders.mockResolvedValue(remoteOk([
+      { provider: 'openai', displayName: 'openai', settingsNs: 'llm-pi-ai', settingsPath: ['providers', 'openai'] },
+    ]))
+    // A pi-ai schema naming no protocol leaves nothing to declare either.
+    const protocolless = JSON.parse(JSON.stringify(
+      Schema.object({ providers: Schema.dict(Schema.object({})) }).toJSON(),
+    )) as JsonValue
+    scripted.face.settings.describe.mockResolvedValue(remoteOk({
+      writable: true, hasDocument: false,
+      namespaces: wireNamespaces().map(view => view.ns === 'llm-pi-ai' ? { ...view, schema: protocolless } : view),
+    }))
+    await mountFace(scripted)
+    expect(screen.getByRole<HTMLButtonElement>('button', { name: en.add }).disabled).toBe(true)
+  })
+
+  it('collapses the add card from the custom mode on cancel', async () => {
+    await mountSection()
+    fireEvent.click(screen.getByRole('button', { name: en.add }))
+    fireEvent.click(screen.getByRole('tab', { name: en.addCustom }))
+    fireEvent.click(within(screen.getByRole('tabpanel', { name: en.addCustom })).getByText(en.cancel))
+    expect(screen.queryByRole('tablist')).toBeNull()
+    expect(screen.getByRole('button', { name: en.add })).toBeTruthy()
+  })
+
   it('cancels the add card back to the add button', async () => {
     await mountSection()
     fireEvent.click(screen.getByText(en.add))
diff --git a/packages/client/ui-settings-models/tests/provider-form.client.spec.tsx b/packages/client/ui-settings-models/tests/provider-form.client.spec.tsx
index 9b6ff1f55a..36dc899778 100644
--- a/packages/client/ui-settings-models/tests/provider-form.client.spec.tsx
+++ b/packages/client/ui-settings-models/tests/provider-form.client.spec.tsx
@@ -225,6 +225,12 @@ function openEditor(provider: string): void {
   fireEvent.click(summary)
 }
 
+/** Open the add card from the section and switch it to the custom-API mode. */
+function openCustomMode(): void {
+  fireEvent.click(screen.getByRole('button', { name: en.add }))
+  fireEvent.click(screen.getByRole('tab', { name: en.addCustom }))
+}
+
 /** Open one model row's advanced fold, where the capacities live. */
 function expandModel(index: number): void {
   fireEvent.click(screen.getByLabelText(`${en.modelAdvanced} ${index}`))
@@ -1417,27 +1423,40 @@ describe('hand-declared providers', () => {
     expect(buttonNamed(en.create).disabled).toBe(true)
   })
 
-  it('closes the create card when an existing row is opened for editing', async () => {
+  it('closes the add card when an existing row is opened for editing', async () => {
     await mountSection({ providers: { openai: { baseURL: 'https://proxy.example/v1' } } })
 
-    fireEvent.click(screen.getByRole('button', { name: en.customAdd }))
-    expect(screen.getByText(en.customTitle)).toBeTruthy()
+    openCustomMode()
+    expect(screen.getByRole('textbox', { name: en.customRoute })).toBeTruthy()
 
     // Two cards at once would each be closable by the other: whichever one is
     // dismissed clears the shared state and discards the other's draft.
     openEditor('openai')
-    expect(screen.queryByText(en.customTitle)).toBeNull()
+    expect(screen.queryByRole('tablist')).toBeNull()
+    expect(screen.queryByRole('textbox', { name: en.customRoute })).toBeNull()
   })
 
-  it('reaches the card from the section and returns to the button on cancel', async () => {
+  it('reaches the custom form from the section and returns to the button on cancel', async () => {
     await mountSection()
 
-    fireEvent.click(screen.getByRole('button', { name: en.customAdd }))
-    expect(screen.getByText(en.customTitle)).toBeTruthy()
+    openCustomMode()
+    expect(screen.getByRole('textbox', { name: en.customRoute })).toBeTruthy()
 
     fireEvent.click(screen.getByText(en.cancel))
-    await waitFor(() => { expect(screen.queryByText(en.customTitle)).toBeNull() })
-    expect(screen.getByRole('button', { name: en.customAdd })).toBeTruthy()
+    await waitFor(() => { expect(screen.queryByRole('textbox', { name: en.customRoute })).toBeNull() })
+    expect(screen.getByRole('button', { name: en.add })).toBeTruthy()
+  })
+
+  it('names each protocol by its product name and falls back to the identifier of an unknown one', () => {
+    mountCard({ protocols: [...PROTOCOLS, 'google-generative-ai'] })
+    const protocol = screen.getByLabelText<HTMLSelectElement>(en.customApi)
+    const labels = [...protocol.options].map(option => [option.value, option.textContent])
+    expect(labels).toEqual([
+      ['openai-completions', en.protocolOpenAiCompletions],
+      ['openai-responses', en.protocolOpenAiResponses],
+      ['anthropic-messages', en.protocolAnthropicMessages],
+      ['google-generative-ai', 'google-generative-ai'],
+    ])
   })
 
   it('refuses an unusable key on the field and blocks creation', () => {
@@ -1608,15 +1627,15 @@ describe('API key field', () => {
     const { controller, mutate } = await mountSection()
     const load = vi.spyOn(controller, 'load')
 
-    fireEvent.click(screen.getByRole('button', { name: en.customAdd }))
-    fireEvent.change(screen.getByLabelText(en.customRoute), { target: { value: 'acme' } })
-    fireEvent.change(screen.getByLabelText(en.baseUrl), { target: { value: 'https://acme.test/v1' } })
+    openCustomMode()
+    fireEvent.change(screen.getByRole('textbox', { name: en.customRoute }), { target: { value: 'acme' } })
+    fireEvent.change(screen.getByRole('textbox', { name: en.baseUrl }), { target: { value: 'https://acme.test/v1' } })
     fireEvent.click(screen.getByRole('button', { name: en.addModel }))
     fireEvent.change(screen.getByLabelText(`${en.modelId} 1`), { target: { value: 'm' } })
     fireEvent.click(screen.getByText(en.create))
 
     await waitFor(() => { expect(mutate).toHaveBeenCalledOnce() })
     await waitFor(() => { expect(load).toHaveBeenCalledOnce() })
-    expect(screen.queryByText(en.customTitle)).toBeNull()
+    expect(screen.queryByRole('textbox', { name: en.customRoute })).toBeNull()
   })
 })