chore(examples/voice/#486): retire legacy gpt-4o-audio-preview surface, migrate supported audio examples to gpt-audio-mini#612
Merged
Conversation
github-actions
Bot
added
the
low-risk-change
![github-actions[bot]](https://avatars.githubusercontent.com/in/15368?s=60&v=4){kind=small}
github-actions
Bot
removed
the
low-risk-change
Collaborator
Author
|
[grinder] READY for human review CI: green (zero failing, zero pending) Verified by: |
drewdrewthis
added
pr-ready
and removed
grinding
Collaborator
Author
✅ Review + prove-it: READY (after closing-ref correction)Review: deleting Prove-it:
Fixed before ready: the body said Minor: carries a disclosed cherry-picked contract-count fix to |
rogeriochaves
previously approved these changes
Jun 5, 2026
…gacy test Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
…t file Companion to the delete commit: `test_voice_to_voice_conversation.py` was removed but two references remained: - docs/scripts/mdx-examples-manifest.js: remove sourcePath entry - .github/workflows/voice-integration.yml: remove from pytest command Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
…ython example The deleted `test_voice_to_voice_conversation.py` was referenced in `docs/docs/pages/examples/multimodal/voice-to-voice.mdx` as: - a generated MDX import (breaking the docs build) - a Python LanguageTabs.CodeTab (now empty) - a prose GitHub link in "Complete Sources" Remove the import, the Python tab, and update the prose link to point to the helper utilities instead with a note about the legacy pattern removal. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
The voice-to-voice example helper and the audio-to-text example pinned
`gpt-4o-audio-preview`, which OpenAI has removed (404 model_not_found
since 2026-05-19). Any user running the canonical voice example hit an
immediate 404.
Switch to `gpt-audio-mini` — OpenAI's current cost-efficient GA
audio-chat model — matching the Python twin, which already migrated
(python/scenario/config/voice_models.py:44 OPENAI_AUDIO_CHAT_MODEL,
python/examples/test_audio_to_text.py:157). Verified live: gpt-audio-mini
accepts the identical chat.completions shape (modalities:["text","audio"],
audio:{voice,format}) and returns audio. Re-ran the voice-to-voice e2e
against prod LangWatch — success: true, real 2-turn conversation, traces
landed (project_bZspxwkhCD4POvqmIgOr2).
SDK core was unaffected (OpenAIRealtimeAgentAdapter uses gpt-realtime-mini).
This closes a py↔ts example-parity gap left by #561.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
… refs after model swap Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
…e, keep+migrate supported audio examples Cohesive retirement of the legacy gpt-4o-audio-preview voice/audio example surface, folding in the model swap from #607 (cherry-picked) and superseding the unskip plan in #486. Genuinely-dead (retired): - Tombstone docs/docs/pages/examples/multimodal/voice-to-voice.mdx and testing-voice-agents.mdx -> pointer to /voice/getting-started (URLs still 200; the langwatch vocs fork has no redirect layer, so a tombstone is how we avoid 404s on previously-public URLs). - Delete the now-unused LegacyVoiceDeprecation.mdx snippet (no importers left). - (test_voice_to_voice_conversation.py already deleted in an earlier #486 commit.) Supported (kept + migrated to gpt-audio-mini): - audio-to-text.mdx / audio-to-audio.mdx kept and updated: prereq prose now names gpt-audio-mini; LegacyVoiceDeprecation banner removed (these document the CURRENT supported single-call pattern, not a legacy one). - Python test_audio_to_text.py / test_audio_to_audio.py: skip COMMENTS rewritten to the real reason (live E2E -- real OpenAI gpt-audio-mini + LangWatch backend, cost, non-deterministic audio); skipif(CI) markers retained by design. No model literal change (they route through the helper's gpt-audio-mini default). - _generated example partials regenerated to match the migrated test sources. overview.mdx voice-agents link repointed to /voice/getting-started. Why the audio tests stay CI-skipped: they are live end-to-end tests; #486's "unskip to restore CI coverage" premise was never achievable (cost + non-determinism). The right end-state is migrated-and-intentionally-skipped. Docs build: pnpm build exits 0, no broken-link/missing-import errors; tombstone routes render the /voice/getting-started pointer. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
drewdrewthis
force-pushed
the
fix/486-delete-deprecated-voice-test
branch
from
4acb9db to
98d970a
Compare
drewdrewthis
changed the title
Collaborator
Author
Review verdict: READYRe-reviewed at HEAD Resolved since last review
Non-blocking
Evidence
|
… symmetric JS twin retirement, drop dead helper exports FIX 1 [blocker]: rewrite specs/voice-docs-surface.feature deprecation group to assert the retirement reality, not the old keep-and-banner strategy. - AC15: reframed to "tombstoned pages still 200, point to /voice/getting-started" and clarifies the supported audio pages stay live+migrated (not tombstoned). - AC17: shared LegacyVoiceDeprecation.mdx snippet is gone; rewritten to an inline per-page tombstone pointer. - AC18: flipped from "source file is not deleted" to assert the file IS deleted and canonical demos live at python/examples/voice/*. Background + AC Coverage Map updated to match. FIX 2 [blocker]: repoint dangling tombstone link in docs/docs/pages/examples/multimodal/multimodal-images.mdx from ./testing-voice-agents to /voice/getting-started (mirrors audio-to-*.mdx). FIX 3 [blocker]: symmetric retirement of the JS voice-to-voice twin — delete the test, its mdx-examples-manifest.js entry, and the orphaned generated mdx. Confirmed no other page imports the generated mdx. FIX 4 [non-blocker]: drop dead helper exports save_conversation_audio + concatenate_wav_files (deleted test was their only consumer; grep shows no other importer) from helpers/__init__.py and their defs in audio_helpers.py. encode_audio_to_base64 is kept (still used by audio-to-audio/text examples). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Contributor
|
Automated low-risk assessment This PR was evaluated against the repository's Low-Risk Pull Requests procedure and does not qualify as low risk.
This PR requires a manual review before merging. |
rogeriochaves
approved these changes
Jun 11, 2026
This was referenced Jun 11, 2026
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
2 participants
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.
What
Cohesive retirement of the legacy
gpt-4o-audio-previewvoice/audio example surface, consolidating three threads into one PR:gpt-4o-audio-preview→gpt-audio-mini) — its commits are cherry-picked here, so fix(examples/voice): swap deleted gpt-4o-audio-preview → gpt-audio-mini #607 is superseded and will be closed.Why we don't unskip (the key correction)
#486 framed the goal as "remove the
skipif(CI)markers and restore CI coverage." That premise was never achievable: these audio/voice example tests are live end-to-end tests — they call real OpenAI (gpt-audio-mini) and the real LangWatch backend, incur cost, and produce non-deterministic audio. They are correctly CI-skipped regardless of model (same class as the other livevoice/*example tests). The deadgpt-4o-audio-previewmodel was only the historical reason for the skip; swapping it doesn't make these CI-runnable.So the right end-state is migrated + intentionally CI-skipped, not "unskipped." We migrate the model so the examples work when run live/locally, keep the skip markers, and fix the stale skip comments to state the real reason.
What changed
Genuinely-dead → retired:
docs/docs/pages/examples/multimodal/voice-to-voice.mdxandtesting-voice-agents.mdx→ a short pointer to/voice/getting-started. The pages' URLs still return 200 (the vocs fork has no redirect layer, so a tombstone avoids 404s on previously-public URLs).LegacyVoiceDeprecation.mdxsnippet (zero importers after the edits).test_voice_to_voice_conversation.pydeleted (it was explicitlyDEPRECATED— the legacy single-call pattern).Supported → kept + migrated to
gpt-audio-mini:audio-to-text.mdx/audio-to-audio.mdxkept and updated (prereq prose now namesgpt-audio-mini;LegacyVoiceDeprecationbanner removed) — these document the current supported single-call pattern, not a legacy one.multimodal-audio-to-text,multimodal-audio-to-audio,multimodal-voice-to-voice-conversation,helpers/openai-voice-agent.ts) migrated togpt-audio-miniwith updated skip-comments (via fix(examples/voice): swap deleted gpt-4o-audio-preview → gpt-audio-mini #607's cherry-picked commits).test_audio_to_text.py/test_audio_to_audio.pyskip-comments rewritten to "live-E2E, not model";skipif(CI)markers retained (no model-literal change — they route through the helper'sgpt-audio-minidefault)._generatedexample partials regenerated to match the migrated test sources.overview.mdxvoice-agents link repointed to/voice/getting-started.Verification
pnpm build(indocs/) exits 0, no broken-link/missing-import errors./voice/getting-startedpointer (confirmed indist/examples/multimodal/voice-to-voice/index.html).audio-to-textbuilt page referencesgpt-audio-mini(x6); no livegpt-4o-audio-previewmodel literal remains anywhere (grepovermodel=/model:is empty — remaining mentions are explanatory comments/tombstone prose only).Closes / supersedes
🤖 Generated with Claude Code