Files
deepseek-harness/packages/experimental/voice-input-bundle
Turtle e7def469e1 feat(i18n): key translation pairing records by heading section (#5036)
* feat(i18n): key translation pairing records by heading section

Replace the whole-file blob hashes in *.i18n.yaml with one entry per
heading section that still holds language-specific content, keyed by the
English heading-slug path. Blocks identical across the pair (generated
table rows, code blocks, shared regions) stay out of the hash, so edits to
different sections and identical regenerations merge with Git's default
text merge, including on GitHub.

Remove the dsh-translation-pairing merge driver, its resolver command and
installer wiring; gen-translation-brief recovers last-confirmed text from
commit history. Migrate all active records, archive the superseded Agent
Note, and record the replayed conflict reduction in the new one.

* test(i18n): refresh translation prompt snapshot for the terminology note

* feat(i18n): exclude only code blocks and generated regions from pairing records

Replace byte-equality masking with an explicit rule: a record hash skips
fenced code blocks and BEGIN/END GENERATED regions, the two kinds of
content the pairing gate already requires to be identical on both sides.

Move generated data into regions written into both pages:
gen-config-catalog renders both languages and the record with per-package
regions labelled by code identifiers; gen-module-graph keeps its table in a
region; gen-doc-graphs splices the event matrix regions into the Chinese
page. Add a generic region splice beside the region grammar.
2026-09-23 22:32:06 +08:00
..
2026-09-23 20:47:45 +08:00

description, kind
description kind
Enable experimental speech input from the plugin manager. package-bundle

@deepseek-ai/dsh-experimental-voice-input-bundle

English | 中文

Summary

This optional bundle composes a speech Service Definition, local SenseVoice provider, authenticated Remote and browser microphone control. Shipped profiles leave it disabled.

Table of Contents


Use this package

Open Plugins in the Web sidebar and enable Voice Input, marked by a blue waveform icon. If models need preparation, a dialog offers Go to setup or Later; Go to setup opens the bundle details. Complete caches need no setup prompt. In the details, choose Download and prepare; the collapsed current-step summary expands to the complete step list. Downloads report actual bytes; verification and loading show elapsed time. Once ready, click the microphone between the model selector and Send, then Stop to insert a transcript. Bundle details store the recognizer and language through the Settings service. Disabling the bundle cancels active work; cached assets remain on disk.


Understand the implementation

Maintainer details — click to expand

The static cordis.patch.yml adds the four voice rows, selects sensevoice-local as the default recognizer and supplies the provider cache directory with dshHomePath. Optional-bundle installation makes the package available to management without selecting it in default profiles. The browser contribution owns its generated Remote mount; stable API Remotes do not import experimental code. No runtime invariant companion is published because this configuration-only package has no independently mutable runtime state.


Further Exploration

Voice input subsystem


Model Experience

None, as recordings and preparation remain outside model requests; ordinary user submission owns any later text.

KV Cache effect

No direct effect; ordinary submission owns the message content.

Known Limitations and Deferred Work

  • The initial bundle supplies one local recognizer. Additional providers register with the same service under distinct ids; cloud recognition requires an explicit new provider and credential configuration. The bundle does not add a model tool or change the agent loop.
  • Installing dsh also installs sherpa-onnx-node and its platform-specific native runtime, including ONNX Runtime, even when this bundle is disabled. Runtime installation adds disk and download costs separate from the models downloaded by Download and prepare; the native package size varies by platform and version.

Dev Note

Maintainer details — click to expand

None.