Bug fixes and broader website support for dictation mode.
Supported websites are still limited for Firefox, but we're working on it.
Let us know which sites matter most to you, and we'll prioritize those!

Say, Pi v1.9.0 - Universal Voice Typing

NEW FEATURE: Voice typing now works on any website, not just Pi and Claude AI chats.

What's new:
- Dictate text into any input field, email, or form across the web
- Same high-quality Whisper transcription as AI conversations
- Real-time text streaming with intelligent punctuation
- Supports 30+ languages with auto-detection

How to use:
- Focus any text field and click the Say, Pi microphone button that appears
- Or right-click any text field and select "Start dictation with Say, Pi"
- Speak naturally, text appears in real-time
- Click button again or pause to finish

This release expands Say, Pi from AI-specific voice chat to universal voice productivity across the entire web.

Note: This is an MVP release. Some websites may require custom optimizations for optimal performance. We're actively improving compatibility based on user feedback, so please let us know where you encounter issues.

This update (v1.8.4) brings a new visual enhancement to your Say, Pi voice chats!

✨ What's New: Transcription Status Indicator ✨

Now, the call button shows you the real-time status of your voice being transcribed:

• See Your Speech Processed: A segmented ring on the call button dynamically updates as you speak.
• Color-Coded Status:
  • Gray: Capturing audio.
  • Blue: Processing to text.
  • Green: Transcribed successfully.
  • Red: Transcription error.
• Clearer Feedback: Get instant visual confirmation of your voice input, making interactions feel more responsive.

We hope you enjoy this improvement to your hands-free experience with Pi and Claude!

Please share any feedback you have with us at info@saypi.ai.

Thanks,
The Say, Pi Team

This update brings a handful of thoughtful improvements focused on enhancing your experience with Claude AI:

🧠 App Activity Tracking
You can now view which assistant you used—Claude or Pi—in your recent activity on the Say, Pi dashboard. A handy way to keep track of your conversations across platforms.

🔊 Smarter Voice Output Filtering
We’ve fine-tuned how Claude’s responses are spoken aloud. The new content filter removes robotic or irrelevant elements (like search result snippets or “thinking…” tokens), so what you hear sounds more like natural conversation.

As always, thanks for using Say, Pi! If you’ve got feedback, we’d love to hear it.

With this release Say, Pi for Claude now:
- works better on mobile 📱
- supports agentic listening 🤖
- lets you know when you're running out of credits 💸
- is generally a bit more stable and polished all round ✨

Say, Pi v1.8.0 - Claude AI Integration

We're excited to announce that Say, Pi now supports Claude AI from Anthropic! This major update expands our voice interaction capabilities beyond Pi AI to bring you a seamless voice experience with Claude.

What's New:

• Claude AI Support: Full voice interaction with Claude AI at claude.ai
• ALTER Voice: A custom-designed premium voice created specifically for Claude's personality
• Mobile Support: Works on Firefox Mobile for on-the-go AI conversations
• Multilingual Support: Same great language capabilities you love with Pi, now for Claude
• Cross-Browser Consistency: Core features work the same across all supported browsers

Note for Firefox Users:
While automatic interruption isn't available in Firefox due to browser limitations, you can use the manual interrupt button to pause the AI when you want to speak - a common approach in many voice assistants.

Getting Started with Claude:
1. Visit claude.ai with Say, Pi installed
2. Look for the call button and press it to start a voice conversation
3. For text-to-speech, sign in through the extension's settings

Thank you for using Say, Pi! We're committed to expanding your voice AI experience across multiple platforms.

Say, Pi v1.7.2 - What's New:

• Improved text-to-speech performance: Paola and Joey voices now respond more quickly and reliably
• Support for 32 languages with improved fluency and natural pronunciation
• Soft launch of "agentic listening" feature for Pro users and beta testers - your AI assistant can now intelligently decide when to speak vs. continue listening, creating more natural conversations

As always, we appreciate your feedback and support as we continue to improve Say, Pi!

New performance visualization tool! Now you can see exactly how long each step takes when talking with Pi.

View a detailed timeline showing speech-to-speech response times
Breakdown includes transcription time, LLM response time, and text-to-speech conversion
Access by clicking the new stopwatch icon in message hover controls
Currently available on desktop only (mobile coming later)

Perfect for our tech-savvy users who want to peek under the hood!

Say, Pi v1.7.0 - Subscriptions & Accounts Are Here!

What's New:
• User Accounts: Track your usage and get ready for future syncing across devices
• Subscription Plans: Choose between Free, Plus, and Pro tiers based on your needs
• Premium Voices Now Live: Our multilingual TTS beta graduates to full release

User Accounts:
• See exactly how you're using Say, Pi with detailed stats
• We're laying the foundation for syncing your settings across devices (coming soon!)
• Don't want an account? No problem! All core Pi AI features work without registration

Choose Your Plan:
• Free: All Pi AI core features remain completely free
• Plus: Higher usage limits + premium multilingual voices (€5.00/month)
• Pro: Maximum usage for power users and professionals (€10.00/month)

Premium Voice Updates:
• Free users get a taste of our premium voices with limited credits
• Plus and Pro subscribers enjoy generous TTS limits in 32 languages
• Check out the full comparison at saypi.ai/pricing

Important Note:
While this update builds the foundation for exciting new features coming soon, Pi AI features remain 100% free. Premium plans unlock enhanced capabilities for Pi AI and will extend to other AI platforms in the future.

- Improved performance of the call button glow animation
- Hardware acceleration support for immersive mode animations
- Removal of redundant code not used in Firefox
- Fixed issue with interrupt button not always working correctly
- Fixed issue where Pi's voice was sometimes recorded after the interrupt button was pressed

- Added a convenient settings button directly in the Pi interface. Access language preferences, voice settings, and other options without leaving Pi. Available in both standard and immersive views. This feature is currently not available for Firefox Android.

- New Features
Enhanced error messaging for call initiation
Improved audio reliability for longer responses

- Known Limitations
Hands-free interrupt unavailable (use tap-to-interrupt instead)
Sound effects temporarily disabled due to echo cancellation constraints