Skip to content

Conversation

@hoffination
Copy link
Owner

Summary

  • Convert the realtime console from conversational AI to a transcription-first translation app
  • Uses Whisper for speech-to-text with auto language detection
  • Translates between two user-selected languages via Chat Completions API (gpt-4o-mini)
  • New timeline UI showing original speech and translations for both speakers
  • Support for VAD (voice activity detection) and push-to-talk input modes

Features

  • Client-side API key: Users provide their own OpenAI key (stored in localStorage)
  • Language pair selection: Choose any two languages from the supported list
  • Collapsible debug sidebar: Toggle to show/hide event log
  • Smart auto-scroll: Timeline scrolls with new messages but pauses when user scrolls up
  • Push-to-talk: Hold button to record, release to transcribe

Test plan

  • Enter API key and verify it persists on refresh
  • Select language pair and start session with VAD mode
  • Speak in one language, verify transcription and translation appear
  • Test push-to-talk mode
  • Verify sidebar collapse/expand works
  • Test auto-scroll behavior (should pause when scrolled up)

🤖 Generated with Claude Code

kwhinnery-openai and others added 26 commits December 19, 2024 16:04
remove outdated reference to Astro and unused Astro files
switch to express server with vite middleware
stop audio capture when session closed
fix issue where timestamps would update on each re-render by adding timestamp to events when they are created/received instead of render time
…istence

fix: preserve original event timestamps
Updating model snapshot from 12-17 to 06-03
Convert the realtime console from a conversational AI demo to a translation app:
- Transcription-only mode using Whisper for speech-to-text
- Auto-detect source language and translate to target language
- Support for VAD (voice activity detection) and push-to-talk input modes
- New ConversationTimeline UI showing spoken text and translations
- Client-side API key input with localStorage persistence
- Collapsible debug sidebar with smart auto-scroll
- Translation via Chat Completions API (gpt-4o-mini)

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude <noreply@anthropic.com>
@vercel
Copy link

vercel bot commented Dec 5, 2025

The latest updates on your projects. Learn more about Vercel for GitHub.

Project Deployment Preview Comments Updated (UTC)
openai-realtime-console Error Error Dec 5, 2025 8:41am

- Remove unused /session endpoint that referenced undefined sessionConfig
- Remove unused apiKey variable from server.js
- Add dismissible error banner UI for user-facing errors
- Improve startSession() with proper try/catch and resource cleanup
- Improve handleTranscription() to check response.ok and handle errors
- Replace alert() calls with error state management

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude <noreply@anthropic.com>
- Update build:server script to use entry-server.jsx instead of index.js
- Remove orphaned client/index.js with invalid framework imports

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude <noreply@anthropic.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

8 participants