Repository navigation
feat(transcription): one speaker label per voice across the parts of an episode - #542
Merged
Merged
Conversation
gpt-4o-transcribe-diarize rejects requests with more than 1400 s of audio, but MP3 chunks were split by size alone. Any MP3 under about 120 kbps put more than 1400 s in a 20 MB chunk, so every chunk failed. For diarization, split MP3 at frame boundaries into the fewest equal parts that stay under both 20 MB and 1400 s. Whisper keeps its size-only chunks. Fixes #532
A chunk that OpenAI rejected with a status such as 400 or 401 was sent three times, although each send fails the same way. A long diarized episode went up three times before the error showed. Retry only the statuses openai-node retries: 408, 409, 429 and 5xx, plus network errors. Refs #532
gpt-4o-transcribe-diarize takes up to 0.58 s per second of audio, measured over 81 requests, so the 1,072 s parts a 54 min episode got would take about 10 min each, at the request timeout. Parts are now at most 600 s, which the model answers in under 6 min. Refs #532
Each test passes on the hand-written code. Together they catch wrong derivations that no existing test sees: the -1 SecretStorage ID suffix that released versions accept, the provider an export error names, the kind a settings key row adopts a secret for, and the key migration notices.
CREDENTIALS in src/types/Credentials.ts holds each kind's settings field, export key and label. The credential repository, settings decoder, settings transfer, settings tab and key migration notice loop over it. IPodNotesSettings takes its key fields from it, so a kind with no default in DEFAULT_SETTINGS fails typecheck. The plaintext key map from before 2.19.0 stays fixed.
The published transcript renderer is the reader of the format that stays when the diarization folder goes. The default template now lives only in DEFAULT_SETTINGS, so src/constants.ts imports no service code.
src/services/transcription/ turns an episode into a transcript with any provider. A provider is one file: its models, its request, and how to read its answers. OpenAI (gpt-transcribe) and Deepgram (nova-3) are the first two. The engine owns what providers share: - audioSource.ts decides whether a provider may be given the episode's address. Only an address that a saved public feed lists, that carries no credential, and that names a public host passes. All other audio is uploaded. - split.ts is audioChunker.ts with the caps as arguments and a start time on each part. - http.ts sends a request at most three times, and only after a 429, a 5xx, or a lost connection. Its failure sentences never include the provider's own text, which can quote a private address. - cues.ts turns words or text into the shapes of a published transcript. Nothing calls the engine yet. Lint rejects an import of the folder's files from outside it, except its index.
TranscriptionService keeps its queue, the published-transcript path, the progress notice, and note saving. It calls the engine for audio. - OpenAI transcribes with gpt-transcribe. whisper-1 and gpt-4o-transcribe-diarize, which OpenAI shuts down on 2027-02-26, are gone. A tooling test searches the built main.js for them. - With speaker labels on and OpenAI selected, the transcript is saved without labels. A notice that stays until clicked names the providers that label speakers. - Deepgram transcripts go through renderTranscriptBody, so each paragraph has a timestamp. Speakers are numbered in the order they first speak. - Deepgram gets the address of a public episode and downloads it itself. - A failure notice stays until clicked. Parts go out one at a time, and a 401, 402, 404, or 400 stops the job before the next part. Deletes src/services/diarization/, openaiTranscription.ts, the Whisper worker pool, and createBinaryChunkFiles. getEpisodeAudioBuffer no longer returns a basename, because uploads are named for their format only.
net.mjs answers OpenAI with { text } and Deepgram with the words of two
speakers, and logs the content type of each request. fixtures.mjs serves a
feed at priv3.supercast.com, an address that PodNotes keeps private. The
transcripts recipe covers OpenAI in parts, and Deepgram by address and by
upload.
The marketplace review test called TranscriptionService.waitForRetry, which the engine replaced. An aborted run now rejects through waitForLifecycle, so the test aborts a pending wait there and expects the same Error.
An aborted job no longer downloads the episode or decodes it into parts. Until now only its requests were refused. A 429 or a 5xx that asks to wait over a minute is sent once. A sooner send would only upload the audio to be refused again.
… feat/transcription-providers
A provider's name was written twice: once in its credential row and once in its provider file. The row now owns it, so a provider file holds only its models and its request. Each model gains the name the settings will show.
The provider was derived from the speaker settings: Deepgram with speaker labels on, OpenAI otherwise. It is now stored as transcript.provider, with transcript.model beside it, where "" follows the provider's recommended model. schemaVersion stays 2 and both fields are additions, so a device on 2.24.0 keeps loading the library and keeps the fields on its own saves. A library that has no transcript.provider gets the provider it ran with: Deepgram for speaker labels with Deepgram, with or without its key, and OpenAI otherwise. The migration reads only the stored document, adds no decode warning, and gives the same answer on every load until a save stores it. transcript.diarization.provider belongs to versions up to 2.24.0. This version reads it for that migration alone, never writes it, keeps it in data.json as found, and leaves it out of settings exports. Importing a 2.24.0 export that labels speakers with Deepgram selects Deepgram. A provider or model ID this version does not have is kept as stored and shown under that name. An unknown provider refuses to transcribe with a notice. An unknown model runs the recommended model and the notice says so. The Transcripts settings now read: Transcription provider, the chosen provider's API key, Model, Speaker labels, Speaker label format, then the path and template. Speaker labels is disabled for a model that cannot label speakers. The Diarization provider row is gone.
ElevenLabs is a credential row, one provider file and a registry line. Its key reference is the new root field elevenLabsSecretId, stored in SecretStorage as podnotes-elevenlabs-api-key, and a settings export carries the key as secrets.elevenLabs. The export envelope stays at version 2. A public episode goes to ElevenLabs by its address, so nothing is downloaded. Any other episode is uploaded in one request. Word times and speakers reach the note, and turning speaker labels off asks for none. ElevenLabs answers 401 for a used-up quota as for a rejected key, and 400 for a model the key may not use as for an address it could not download, so the provider reads detail.code. Only the last of these is followed by an upload. A request waits 10 minutes for its answer. A new install starts on ElevenLabs with speaker labels on. A library saved up to 2.24.0 that holds no OpenAI key and has speaker labels off moves to ElevenLabs too, because it never set transcription up. A library that never stored the speaker labels choice keeps them off. The notice for a model without speaker labels now names ElevenLabs first.
A journey now seeds its transcription provider with its key, because a new vault is on ElevenLabs. The OpenAI journeys choose OpenAI. A new journey runs the default: a public episode goes to ElevenLabs by its address, no audio is fetched, and the note has timestamps and speaker labels. The ElevenLabs stub answers with the words ElevenLabs returns and records what each request named. The settings tab test walks the new Transcripts rows, and the key recovery test links the ElevenLabs key, the one a new vault shows.
The fixture provider answers api.elevenlabs.io with ElevenLabs' words, two speakers with the second 12 s in, and logs the address a request named in place of a file. The transcript guide follows the new settings rows and adds the ElevenLabs runs and a 2.24.0 library check.
…choice The provider row decided for itself whether the stored provider is one this version has, and the Speaker label format row read the stored speaker setting. Both now ask transcriptionChoice, as the key rows, the model row and the Speaker labels row do. The engine no longer exports a type nothing imports.
The Transcripts tests looked each control up by a CSS selector inside its row and asserted on a hidden row's inline style. They now query by role and by the row's name, as a user or a screen reader finds a control, so a hidden row is one the query does not find.
…ings measure The transcription engine left main.js at 103,795 gzip bytes. The provider and model settings, their migration and the ElevenLabs provider add 866, which takes it to 104,661 and over the 104,000 budget. The budget rises by those 866 bytes to 104,866, so the 205 bytes of headroom stay as they were.
…m library 2.24.0 imports the diarization object whole and fills a missing provider with OpenAI, so omitting the field moved Deepgram libraries to OpenAI. Imports here keep the provider this library stored.
A library migrated with OpenAI speaker labels on could not turn them off, so every transcription repeated the notice.
… one The tests for a provider and a model a newer PodNotes stored used OpenRouter and Grok STT 2.0, which this version is about to run. They now use AssemblyAI and universal-3-6-pro.
OpenRouter is a credential row, one provider file and a registry line. Its key reference is the new root field openRouterSecretId, stored in SecretStorage as podnotes-openrouter-api-key, and a settings export carries the key as secrets.openRouter. It offers three models, each with the request measured to give word times and speakers: - Grok STT 2.0, recommended: base64 JSON with verbose_json and word and segment times, routed to xAI with provider.only, which takes diarize from provider.options.xai. A request routed elsewhere is answered with no speakers. - ElevenLabs Scribe v2: multipart with diarize beside the model, or JSON with input_audio.url for an episode a saved public feed lists. - MAI-Transcribe 2: multipart with diarize beside the model. With speaker labels off, no request carries diarize. Sounds Scribe v2 lists as audio_event items are left out. OpenRouter takes 25 MB a request and gives the model 60 s, so an upload goes in parts of at most 18 MB and 1,800 s. Parts of 24 MB were answered in 20 to 31 s. A 402 says the account is out of credit, a 401 that the key was rejected, and a 404 that the model is not offered. A 404 whose message names the account's data policy or zero data retention says the model is not offered under the account's OpenRouter privacy settings. Each ends the job at the first part, and none quotes OpenRouter.
…t each part A note saved with speaker labels asked for and none returned now says so: "Saved without speaker labels, because Grok STT 2.0 returned none." OpenRouter answers 200 with no speakers when a request reaches a route that ignores the speaker option, so this is checked for every provider. Labels from two parts already name different speakers. An episode sent in parts now also says that its labels restart in each part, so one person can have several labels, until labels are merged across parts.
…ll lists PodNotes reads OpenRouter's public catalog of transcription models, without the key, when the Model row draws and when an OpenRouter job starts. It keeps the read in memory for 24 hours. A read that fails is not kept, so the next one asks again, and the tested list is offered as it is. The catalog only takes models away. A model it does not list is not offered, and one whose expiration_date has come is skipped. A model with a later end date reads "Grok STT 2.0, ends <date>" in the Model row. A catalog that would leave no model is not followed. A chosen model the catalog no longer lists stays selected in the Model row as "<label>, no longer offered". A job runs the first offered model instead and says "OpenRouter no longer offers ElevenLabs Scribe v2, so PodNotes used Grok STT 2.0." The row redraws when a read lands while it is open. A provider gets its catalog through an optional catalog() beside failure(), so the registry owns the cache and the filter for any provider that has one.
A stub answers OpenRouter's transcriptions with two speakers, the second 12 s in, and its model catalog with the three models PodNotes offers. It records what each request named and a hash of the audio it carried, decoded from base64 for JSON. J-TR-09 plays a public episode on Scribe v2: one JSON request names the episode's address, and the episode's audio is never fetched. J-TR-10 plays it on the recommended Grok STT 2.0: one JSON request carries the audio as base64, the same bytes the fixture server holds. Both notes have timestamps and numbered speakers, and the catalog is read once, with no Authorization. A journey seeds any provider by its credential row, and a model with it.
…eads them The fixture provider answers openrouter.ai transcriptions with verbose_json words from two speakers, the second 12 s in, and the model catalog with the three models PodNotes offers. The log keeps openrouter.ai paths and reads the address a JSON request names in input_audio.url. The transcript guide adds the OpenRouter rows, its runs (Scribe v2 by address, Grok by base64, a long MP3 in parts) and the notices that now name OpenRouter.
Frame-like bytes inside a leading tag set the first frame's duration, and counting every part in that frame's length sent a 1,728 s episode of 72 ms frames up in one request. Parts now fit by the summed duration of their own frames, and each part's start is the sum of the frames before it.
Audio frames inside a tag's declared size were counted and could be cut through, sending an undecodable first part, and frame-like bytes in a trailing tag added a part.
…engine # Conflicts: # src/services/TranscriptionService.test.ts # src/services/TranscriptionService.ts # src/services/audioChunker.test.ts # src/services/audioChunker.ts # src/services/diarization/openaiProvider.ts # src/services/diarization/types.ts # src/services/openaiTranscription.test.ts # src/services/openaiTranscription.ts # tests/mocks/openaiFailures.ts
…h cap A file in a format other than MP3 under the size cap went whole whatever its length, so a long .m4a passed a model's maxSeconds. Its length now comes from its metadata through a media element, without decoding it, and audio over the cap is decoded into WAV parts that fit both caps. Audio of an unknown length still goes whole.
A public feed's enclosure with AWS or Google Cloud signing parameters passed the feed credential check, so Deepgram got the whole signed address. A query parameter whose name holds signature, credential or token, or Akamai's hdnts, now uploads the audio instead.
…nload A WebView may load no metadata for media it does not play, so the length probe could wait forever, and an unload could not stop it before a decode. The probe now reads the length as unknown after 5 s or when the job is aborted, and an aborted job decodes nothing.
…nscription-providers # Conflicts: # AGENTS.md # scripts/size-baseline.md # scripts/size.mjs # scripts/size.test.ts # src/persistence/settingsCodec.ts # src/settingsTransfer.ts # src/types/Credentials.ts # src/ui/settings/PodNotesSettingsTab.ts
…nscription-openrouter # Conflicts: # AGENTS.md # scripts/size-baseline.md # scripts/size.mjs # scripts/size.test.ts
…address An address goes in one request, so Scribe v2 on OpenRouter got a public episode of any length by address, past the 1,800 s its parts are capped at for OpenRouter's 60 s limit. An episode the feed says is longer than the model's cap is now uploaded in parts. One of unknown length still goes by address.
…nscription-speaker-merge # Conflicts: # scripts/size-baseline.md # scripts/size.mjs # scripts/size.test.ts
Owner
Author
|
Retargeted to
|
Owner
Author
|
FAIL
|
One speaker's words across a long silence made one turn, so the reel could clip the silence instead of the voice, and that voice kept two labels.
Owner
Author
|
PASS
|
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
OpenRouter takes at most 18 MB and 1,800 s a request, so PodNotes uploads a long episode in parts, and each part numbers its speakers on its own. On the 54 min E231 in 3 parts, those labels agreed with a whole-file run 0.344 of the time. With this change, Grok STT 2.0 gives each voice one number across all parts of an episode. Builds on #540.
mergeSpeakerson the model profile turns the reel on. It is on forx-ai/grok-stt-2.0and off forelevenlabs/scribe-v2andmicrosoft/mai-transcribe-2.Speaker accuracy of the lab implementation (
diarize-lab/scripts/step0.mjs) through OpenRouter:Scribe v2's reel heard Planet Money's 10 voices as 4 (0.378 to 0.414), so it stays off. MAI-Transcribe 2 is not measured. This PR's code has not run against the live API, because this session has no key.
Verification
npm run testpasses 2,888 tests. The new tests send MP3 fixtures whose frames carry a voice and a frame index through therequestUrlseam. The mock provider hears each frame's voice and numbers voices in the order they speak in each request, as a diarizing model does. On the base, the 4 tests for new behavior fail, for example with1 Ann, 2 Bob, 3 Bob, 4 Annwhere1 Ann, 2 Bob, 2 Bob, 1 Annis expected.1 failed) and passed once restored (1 passed), throughnpm run plant:3 Bobfor2 Bob)2 Cyfor3 Cy)lint,format:check,typecheck,build,test:tooling(48 passed) anddocs:buildpass.npm run test:e2e -- tests/e2e/journey-resilience.test.ts tests/e2e/journey-transcript.test.tspasses 26 tests.npm run sizereports 106,873 gzip bytes. The budget rises by the measured 808 bytes, against 106,065 for the base built the same way, and keeps the base's 97 bytes of headroom.Fixtures/Two Voices.mp3is 70 min of 8 kbps, 8 kHz mono, with the two tones taking turns every 45 s. With Grok and labels on,Transcribe current episode from audiosent three part POSTs of about 1.87 MB each, then one reel POST of 82,291 bytes that Chromium decoded to 61.56 s, which is six 9 s clips and five gaps of 21 silent frames. The second part starts with the high tone and numbers the voices the other way. The note still numbers all 47 low-voice paragraphs1and all 47 high-voice paragraphs2, and the paragraph at each part boundary runs on across it. No notice stayed. On Scribe v2 the same file went as three multipart POSTs with no reel, the note numbered six speakers, and the restart warning stayed.dev:errorsreported none.Real OpenRouter key (lead, after review)
Obsidian 1.14.4, Linux, build of
5eebf37b, Grok STT 2.0 (recommended), speaker labels on, the 54 min, 39 MB local E231 interview (two voices), palette command, a pass-through request log:Transcription completed and saved., no restart warning,dev:errorsnone.Release impact
None for settings, storage, the API or URIs. A Grok transcript of an episode over 18 MB or 1,800 s costs one more OpenRouter request of about a minute of audio. OpenRouter is unreleased, so no released behavior changes.
Note
Add speaker-reel matching to merge speaker labels across transcription parts
mergeSpeakers: true(Grok STT 2.0 only). The reel is built from each speaker's longest turn (≥2.6s, trimmed to 9s) and the response maps part-specific labels to one shared number in transcribe.ts and cues.tscutReelassembles MP3 clips with generated silent frames between themTranscriptionError, part transcripts still return and unmapped labels keep the existing label-restart warningMacroscope summarized 5958579.
Note
Medium Risk
Adds an extra provider request and MP3 slicing on multi-part Grok jobs; failures degrade to prior per-part labels rather than failing the job.
Overview
Multi-part OpenRouter transcriptions can now keep one speaker number per voice across the whole episode when the model supports it (enabled for Grok STT 2.0 only via
mergeSpeakers).After all parts finish, the pipeline may send an extra speaker reel: a short MP3 built from each part-local speaker’s longest turn (≥2.6 s, trimmed, clipped at frame boundaries, separated by 1.5 s silence). The reel response is mapped back to part labels through
speakerKey/SameSpeaker, andtranscriptFromPartsapplies that mapping when numbering cues. Scribe v2 and MAI-Transcribe 2 keepmergeSpeakers: falsebecause reel matching was unreliable in measurement. A failed reel still saves the transcript and falls back to the existing “labels restart in each part” warning only for labels that were not merged.Verify tooling’s fixture provider decodes Grok MP3 uploads (tone-based diarization) so e2e can assert reel requests. Bundle gzip budget rises to 107,399 bytes for the new reel code.
Reviewed by Cursor Bugbot for commit 5958579. Bugbot is set up for automated code reviews on this repo. Configure here.