vault/vrds
2
0
Fork 0
Code Issues 3 Pull requests Projects Releases 147 Packages Wiki Activity Actions

feat!: use OpenAI-compatible voice transcription #154

Merged
Wuast94 merged 1 commit from wuast94/openai-compatible-voice-transcription into main 2026-07-03 13:33:17 +02:00 AGit
Conversation 0 Commits 1 Files changed 7 +453 -30
Wuast94 commented 2026-07-03 11:14:10 +02:00
Owner
Copy link

Summary\n- Switch voice transcription to an OpenAI-compatible /audio/transcriptions client with configurable base URL, model, language, and optional bearer API key.\n- Default raw transcription language to German (de) while keeping the existing long-message summarization flow unchanged.\n- Keep legacy modules.voice.whisper_url loading for migration, and mask/preserve the new voice provider API key in runtime config.\n\n## BREAKING CHANGE\nVoice transcription configuration now uses modules.voice.base_url, model, language, and api_key for the OpenAI-compatible provider shape. Existing whisper_url still deserializes for migration, but the web/runtime defaults and preferred config shape changed.\n\nFor the local OVMS deployment without auth, leave modules.voice.api_key empty.\n\n## Test Plan\n- just format\n- just lint\n- just test\n- just everything\n- WHISPER_URL=http://host.docker.internal:18000/v3/audio/transcriptions cargo test test_e2e_transcription_wav --all-features -- --ignored against the live ovms/ovms-whisper service via kubectl port-forward\n

## Summary\n- Switch voice transcription to an OpenAI-compatible `/audio/transcriptions` client with configurable base URL, model, language, and optional bearer API key.\n- Default raw transcription language to German (`de`) while keeping the existing long-message summarization flow unchanged.\n- Keep legacy `modules.voice.whisper_url` loading for migration, and mask/preserve the new voice provider API key in runtime config.\n\n## BREAKING CHANGE\nVoice transcription configuration now uses `modules.voice.base_url`, `model`, `language`, and `api_key` for the OpenAI-compatible provider shape. Existing `whisper_url` still deserializes for migration, but the web/runtime defaults and preferred config shape changed.\n\nFor the local OVMS deployment without auth, leave `modules.voice.api_key` empty.\n\n## Test Plan\n- `just format`\n- `just lint`\n- `just test`\n- `just everything`\n- `WHISPER_URL=http://host.docker.internal:18000/v3/audio/transcriptions cargo test test_e2e_transcription_wav --all-features -- --ignored` against the live `ovms/ovms-whisper` service via `kubectl port-forward`\n
feat!: use OpenAI-compatible voice transcription
Some checks failed
CI / Build (pull_request) Has been cancelled
Details
CI / SBOM (pull_request) Has been cancelled
Details
CI / Quality (pull_request) Has been cancelled
Details
cfd2ffd0f3
 
BREAKING CHANGE: voice transcription config now uses OpenAI-compatible base_url/model/language/api_key fields; legacy whisper_url still deserializes for migration.
Wuast94 force-pushed wuast94/openai-compatible-voice-transcription from cfd2ffd0f3
Some checks failed
CI / Build (pull_request) Has been cancelled
Details
CI / SBOM (pull_request) Has been cancelled
Details
CI / Quality (pull_request) Has been cancelled
Details
to ccd200b3de
Some checks failed
CI / SBOM (pull_request) Failing after 2m20s
Details
CI / Quality (pull_request) Successful in 7m10s
Details
CI / Build (pull_request) Successful in 56m23s
Details
2026-07-03 11:15:13 +02:00 Compare
Sign in to join this conversation.
Reviewers
No reviewers
Labels
Clear labels
No items
No labels
Milestone
Clear milestone
No items
No milestone
Projects
Clear projects
No items
No project
Assignees
Clear assignees
No assignees
1 participant
Notifications
Due date
The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Dependencies

No dependencies set.

Reference
vault/vrds!154
No description provided.