Universal Dictation Platform Support Roadmap

[rosscado]

A prioritised list of website to support for voice typing/dictation.

🎯 Phase 1: Quick Wins

• WordPress - Add auto-button appearance (works via right-click, need auto-focus trigger)
• Groq Chat - Input is via a standard textarea, but an iframe is getting in the way

🚀 Phase 2: High-Impact Platforms

• Facebook - Fix transcript writing to fields (audio converts but text not written)
• Notion - Handle content-editable div (transcripts not presented normally)
• WhatsApp Web - Content-editable div support (audio transcribed but no text display)
• Discord Web - Content-editable div support (audio transcribed and text displayed, but not usably)
• Perplexity - Audio converts but text not written to field. Also, interferes with native dictation mode. Suspected CORS issue.

🔬 Phase 3: Complex Platforms

• Reddit - Custom editor integration (<shreddit-composer/> element)
• Medium - Fix text placement (transcript appears in wrong location)
• Google Docs - Custom editor integration (non-standard text editor element)
• Venice Chat - Input via standard textarea (like Groq), and largely works well apart from some initial button positioning issues, but gets confused by conversational state (turn taking clearing the prompt). Need to handle text deletions.

---

Technical Notes

• These sites work great: Gmail ✅, Slack ✅, Twitter ✅, WriterDuet ✅, GitHub ✅, ChatGPT ✅, Mistral (Le Chat) ✅, Gemini AI ✅, Character AI ✅
• Main challenge: Content-editable divs (Notion, WhatsApp, Medium, Reddit)
• Quick fix pattern: Nested input detection (WordPress)

Success Criteria

• Reddit community can advocate for Say, Pi
• Major social platforms supported (Facebook)
• Productivity workflows enabled (Notion)
• Global messaging coverage (WhatsApp)