feat(models): add Gemini TTS on Vertex+AI Studio

[steebchen]

Summary

Adds the two Gemini text-to-speech models via both Google Vertex AI and Google AI Studio:

• Gemini 3.1 Flash TTS Preview (gemini-3.1-flash-tts-preview, released 2026-04-15): new model entry with google-ai-studio and google-vertex mappings — $1.00/M text input, $20.00/M audio output (per Gemini API pricing), 8k context / 16k max output, same 30-voice catalog as the 2.5 TTS family.
• Gemini 2.5 Pro TTS: adds a google-vertex mapping to the existing gemini-2.5-pro-preview-tts model using the GA Vertex model id gemini-2.5-pro-tts (the preview id is AI-Studio-only, per the Gemini-TTS docs). Same pricing as the existing AI Studio mapping ($1/M in, $20/M audio out). Both models are served on Vertex's global location, matching the gateway's default Vertex region.

Gateway: google-vertex support for /v1/audio/speech

The speech endpoint previously only supported google-ai-studio, openai, and elevenlabs — a Vertex TTS mapping would have been rejected with 400 unsupported_provider. This PR adds a google-vertex branch mirroring the embeddings Vertex path:

• URL: https://aiplatform.googleapis.com/v1/projects/{project}/locations/{region}/publishers/google/models/{model}:generateContent
• Project id from google_vertex_project_id provider-key option or LLM_GOOGLE_CLOUD_PROJECT; region from LLM_GOOGLE_VERTEX_REGION defaulting to global
• Token type via resolveVertexTokenType so the Authorization: Bearer header and the ?key= query param always agree; the resolved type is passed to getProviderHeaders
• Response parsing is unchanged — Vertex returns the same inline-PCM generateContent shape as AI Studio

Tests

• New speech.e2e.ts driven by a new speechModels list in chat-helpers.e2e.ts (same TEST_MODELS/deactivation/stability filtering as embeddingModels), asserting a valid RIFF/WAVE payload. The speech endpoint previously had no e2e coverage.
• Unit specs: two new speech.spec.ts tests covering the Vertex WAV path (including usedModelMapping/cost assertions) and the 3.1 Flash PCM path; the mock server's Vertex generateContent route now handles the AUDIO response modality.
• Live e2e run (real providers): all four mappings return real audio (215–290 KB WAV):

  TEST_MODELS="google-ai-studio/gemini-3.1-flash-tts-preview,google-vertex/gemini-3.1-flash-tts-preview,google-vertex/gemini-2.5-pro-preview-tts,google-ai-studio/gemini-2.5-pro-preview-tts" pnpm test:e2e
  ✓ speech google-ai-studio/gemini-2.5-pro-preview-tts
  ✓ speech google-vertex/gemini-2.5-pro-preview-tts
  ✓ speech google-ai-studio/gemini-3.1-flash-tts-preview
  ✓ speech google-vertex/gemini-3.1-flash-tts-preview
  

• Full pnpm test:unit (171 files, 2852 tests) and pnpm build pass; pnpm format run.

🤖 Generated with Claude Code

[chatgpt-codex-connector]

You have reached your Codex usage limits for code reviews. You can see your limits in the Codex usage dashboard.

[Copilot]

Copilot was unable to review this pull request because the user who requested the review has reached their quota limit.

[coderabbitai]

Warning

Review limit reached

@steebchen, you've reached your PR review limit, so we couldn't start this review.

Next review available in: 13 minutes

Enable usage-based reviews in Billing to review now. Otherwise, wait until the next included review is available.
You're only billed for reviews past your plan's rate limits ($0.25/file).

How can I continue?

After more reviews become available, a review can be triggered using the @coderabbitai review command as a PR comment. Alternatively, push new commits to this PR.

To avoid repeated limits, reduce automatic review volume by pausing incremental auto-reviews earlier, using label-based review opt-in, excluding WIP or generated PR titles, or requesting reviews manually when the PR is ready. If your team needs uninterrupted high-volume reviews, an organization admin can enable usage-based reviews.

How do review limits work?

CodeRabbit enforces per-developer PR review limits for each organization. Most developers receive the normal plan review availability.

For paid Pro and Pro+ PR reviews, CodeRabbit uses adaptive limits for sustained high-volume activity. When a developer's recent PR review activity reaches the 95th percentile or higher among CodeRabbit users, additional reviews become available more gradually as earlier reviews age out of the rolling window.

Please refer docs for additional details.

Review details

⚙️ Run configuration

Configuration used: Repository UI

Review profile: CHILL

Plan: Pro

Run ID: 08c61cb8-b687-48b7-b53f-b2e13f86fdc3

📥 Commits

Reviewing files that changed from the base of the PR and between 135a701 and b52ee34.

📒 Files selected for processing (6)

• apps/gateway/src/chat-helpers.e2e.ts
• apps/gateway/src/speech.e2e.ts
• apps/gateway/src/speech/speech.spec.ts
• apps/gateway/src/speech/speech.ts
• apps/gateway/src/test-utils/mock-openai-server.ts
• packages/models/src/models/google.ts

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Commit unit tests in branch add-gemini-tts-models

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}