feat(pipeline): stage Speechify Simba 3.2 (transport verified live)#5
Open
shauntrennery wants to merge 1 commit into
Open
feat(pipeline): stage Speechify Simba 3.2 (transport verified live)#5shauntrennery wants to merge 1 commit into
shauntrennery wants to merge 1 commit into
Conversation
…egistration blocked on clone-model support) Stages everything a contributor can per docs/ADDING_A_MODEL.md and CONTRIBUTING.md; the catalog entry is intentionally NOT added (clips must be hosted and verified before a registry change lands, maintainers only). - transports/speechify.ts: synthesis (POST /v1/audio/speech, base64 audio_data -> mp3), cloning (multipart POST /v1/voices with consent declaration), TTFB plan (chunked HTTP /v1/audio/stream). All three surfaces verified against the live API on 2026-07-17, including a clone create + synth + delete round-trip on a throwaway voice. - models.ts: pre-registration row speechify-simba-3-2 (proposed frozen arenaApiId simba-3-2, vendor id simba-3.2). - voices.ts: empty speechify clone-slot; .env.example: SPEECHIFY_API_KEY plus the clone consent vars. - RUNBOOK: status-table row + provider quirks (strict model validation, 429 pacing, base64 response shape). Blocker for registration: cloned voices list only simba-english / simba-multilingual today; simba-3.2 is limited to eight stock *_32 shared voices, so the same-voice arena methodology cannot run until Speechify enables clones for it. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
|
@shauntrennery is attempting to deploy a commit to the Vapi Team on Vercel. A member of the Team first needs to authorize it. |
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
1 participant
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.
Stages Speechify Simba 3.2 support up to the contributor line drawn by
docs/ADDING_A_MODEL.mdandCONTRIBUTING.md: the transport and pipeline wiring land here; the catalog registry entry is intentionally not included (clips must be hosted and verified before a registry change lands, and that requires the maintainers' licensed source voices). Proposal issue: #4.What's here
src/pipeline/transports/speechify.ts— fullProviderTransport:synthesize:POST /v1/audio/speech, base64audio_datadecoded to mp3.createClone: multipartPOST /v1/voiceswith the consent declaration (SPEECHIFY_CONSENT_FULL_NAME/SPEECHIFY_CONSENT_EMAIL, added to.env.example).ttfbPlanFor: chunked HTTPPOST /v1/audio/streamviahttpStreamTrial, ready for the 50-trial bench.src/pipeline/models.ts— pre-registration rowspeechify-simba-3-2(proposed frozen ids: slugspeechify-simba-3-2,arenaApiId: simba-3-2following the dotless Inworld convention, vendor idsimba-3.2).src/pipeline/voices.ts— emptyspeechifyclone slot forhumanness:clone/voices.local.json.src/pipeline/RUNBOOK.md— status-table row + provider quirks (strictmodelvalidation, 429 pacing, base64 response shape, bench-voice note).Live verification (2026-07-17)
Every transport surface was exercised against
api.speechify.aiwith a real key (kept in the gitignoredpipeline/.env):simba-3.2x stock voiceharper_32: HTTP 200, valid mp3.simba-english,simba-multilingual,simba-3.0,simba-3.2) — no shared-pool risk.modelslist, deleted it (204). No state left on the account./v1/audio/stream(single observation, not a bench; latency stays measured-only per house rules).Why registration is blocked (and staged anyway)
Cloned voices currently list only
simba-english/simba-multilingual; clone xsimba-3.2synthesis returns 400. Simba 3.2 works only with eight stock*_32shared voices today, so the same-voice methodology cannot run. The RUNBOOK row documents the re-probe (create a clone, check whethermodelsincludessimba-3.2); when Speechify lifts the gate, registration is the standard maintainer steps 1-5 with zero extra wiring.Checklist (CONTRIBUTING.md)
bun run check-typescleanbun testgreen (109/109)bun run buildsucceeds🤖 Generated with Claude Code