Say, Pi v1.5.1 Changelog (GreasyFork Edition)

Hello Say, Pi users! We're excited to bring you version 1.5.1 with some nifty enhancements to improve your experience. Here’s what’s new:

1. Audible Cues for Interaction: Now, you’ll hear a subtle sound at the start and end of your calls, and when the app switches from listening to responding. These cues are designed to keep you informed about what Say, Pi is doing without needing to look at your screen.

2. Progress Ring for Listening Time: Wondering how much time you have left to speak? We’ve added a visual timer around the call button. This progress ring gives you a clear view of how much time is left in the listening phase, so you can pace your speech comfortably.

3. Glowing Call Button: During a call, the call button will now glow gently when Say, Pi is actively listening to you. This not only makes the interface more lively but also provides a clear visual cue for the listening phase.

What hasn't changed:

• Your privacy and security are still our top priority. There are no new permissions or changes in how we handle your data.

As always, we’re keen to hear your feedback. If you encounter any issues or have suggestions for future updates, please let us know!

Happy conversing with Say, Pi!

Changelog for 'Say, Pi' Userscript - Version 1.5.0

Enhancements:

• Speed Improvements: We've optimized Say, Pi for faster performance, reducing latency and enhancing the overall speed of transcription.
• Stability Enhancements: Numerous stability improvements ensure a smoother and more reliable transcription experience.
• General Optimizations: Under-the-hood tweaks have been made to improve efficiency and responsiveness of the script.

Fixes:

• Bug Fixes: Addressed various bugs reported by our users, leading to a more seamless transcription process.

Notes:

• In this update, we've focused on enhancing the core functionality and performance of transcriptions. Users will notice improvements in speed and stability during transcription, with reduced latency and fewer errors.
• Our Chrome and Edge browser extensions offer the option to set a preferred transcription mode that is not available in the userscript version. Get the extension from the Chrome Web Store if you'd like to try that feature. Userscript users will still benefit from the backend improvements that contribute to the overall efficiency of the service.

How to Update:

• To update to the latest version, visit the GreasyFork page for 'Say, Pi' and install the updated userscript.

We're committed to continuously improving 'Say, Pi', and we appreciate all the feedback from our user community. Keep sharing your experiences and suggestions with us!

Say, Pi v1.4.9: Now Available as a Google Chrome Extension!

We are thrilled to announce the release of "Say, Pi" v1.4.9, which marks an exciting milestone: the introduction of "Say, Pi" as a fully-fledged Google Chrome extension! This release brings the seamless speech-to-text capabilities of "Say, Pi" directly into your browser, enhancing your experience with the Inflection AI's Pi chatbot.

What's New in v1.4.9:

• Chrome Extension Launch: "Say, Pi" is now available as a Chrome extension, offering a more stable and integrated experience.
• Enhanced Performance: We've fine-tuned the performance, ensuring faster load times and more responsive interactions.
• Optimised Audio Processing: The audio processing pipeline has been optimised for the Chrome environment, providing high-accuracy transcription in real-time.
• Local WASM File Loading: We've shifted to loading WebAssembly (WASM) files locally within the extension, enhancing security and reducing dependency on external resources.
• Refined User Interface: The user interface has been polished to integrate seamlessly with the Chrome browser, providing a more intuitive and cohesive experience.

Key Features:

• Hands-Free Interaction: Enjoy hands-free conversations with Pi, powered by accurate and real-time speech-to-text transcription.
• Multi-Language Support: Communicate in any language, with high-accuracy transcription across various languages.
• Easy to Install and Use: Simply add the extension to your Chrome browser and start conversing with Pi using your voice.

Installation:

The extension is available on the Chrome Web Store for all Chromium-based web browsers, and will continue to be available as a userscript from GreasyFork and elsewhere too.

Say, Pi v1.4.8 Changelog:

• New Conversational Tempo Feature: Say, Pi now adapts to how fast or slow you speak, providing responses at a pace that feels right for your conversation style.
• Smarter Response Timing: Enhanced timing algorithms mean Say, Pi is more responsive to your speech patterns, ensuring a smoother conversation.
• Quality Improvements: Various improvements have been made to enhance the stability and performance of Say, Pi.

Update to the latest version to experience these improvements. We value your feedback, so please share your experience with the new features!

New in v1.4.7:

• Achieved a notable enhancement in transcription accuracy by providing additional context to the transcription engine. This clever tweak makes your interactions with Say, Pi even more precise than before, quite a feat given the already high level of accuracy Say, Pi has been delivering. Your conversations are now understood with an extra touch of precision!

Say, Pi v1.4.6 – The Compact Yet Robust Update Now on Greasy Fork!

Introduction:

Unveiling Say, Pi v1.4.6, the much-anticipated update now released on GreasyFork.org! This release leapfrogs from v1.3.8 straight to v1.4.6 on Greasy Fork, bridging the gap for users who missed out on the intermediate versions (v1.4.0 to v1.4.5) due to size constraints on userscript repositories. Fret not, as these versions were available on our official website, www.saypi.ai. However, we recognised that our Asian user base, who predominantly installed Say, Pi from userscript repositories, were left in the dark. Hence, we're bringing the latest goodies right back to your favourite repo!

What’s New?

Version 1.4.6 is a packaging marvel, slashing the userscript bundle size by an astonishing 86%, making it repository-friendly once again. While there's no additional functionality over v1.4.5, the size optimisation is the hero here, ensuring the script finds its rightful place on Greasy Fork and other popular userscript repositories.

If you've been on the v1.3.8 train until now, prepare to be swept off your feet by the sweeping enhancements brought in since v1.4.0:

1. Fully Hands-Free Interaction:

   • Experience the liberating hands-free capability that has been the hallmark of Say, Pi since v1.4.0. Your voice is now your command!

2. Class-Leading End-of-Speech Detection:

   • Say goodbye to awkward silences. Our refined end-of-speech detection knows precisely when you've finished speaking, making interactions smooth and natural.

3. Swift and Accurate Speech Recognition:

   • Powered by OpenAI's Whisper API, the speech recognition is both speedy and precise, ensuring your commands are understood in a flash!

How to Get It?

Head over to Greasy Fork - Say, Pi v1.4.6 to grab the latest version and enjoy a more streamlined, repository-friendly Say, Pi. For those who prefer direct downloads, fret not, www.saypi.ai continues to host the latest and greatest versions of Say, Pi.

Support and Feedback:

We value your feedback and are here to assist with any queries or issues. Reach out to us through the #Say, Pi post on the Pi Party Discord server, the feedback form on this page, or on Twitter/X (@saypi_ai) to share your experiences and interact with other Say, Pi enthusiasts.

Conclusion:

This update is all about bringing the phenomenal hands-free features of Say, Pi back to the Greasy Fork community. With a significantly optimised package size, we've ensured no one misses out on the remarkable enhancements Say, Pi has to offer. Happy voice commanding!

:mega: Say, Pi Version 1.3.8: A Focus on Safari Compatibility 🦁

Hey awesome people! We have a minor but important update rolling out, especially aimed at our Safari users.

:toolbox: What's New in v1.3.8:

• :globe_with_meridians: Enhanced Safari Support: For all you Safari aficionados, I've made tweaks to better handle media autoplay. Now you'll see a "play" button in place of the previous "bounce" animation, making it clearer when you need to take action for Pi to speak.

• :tools: General Fixes: While the spotlight is on Safari, this update also includes tweaks and fixes that benefit all browsers.

:loudspeaker: What You Need To Do:

• :arrows_counterclockwise: Upgrading is advised for all users, even if you're not on Safari, but there's no mad rush. Your userscript manager should automatically update Say, Pi to v1.3.8 soon.

:woman_shrugging: Note: I must admit, the fix for Safari still has some quirks. It's less seamless than other browsers due to Safari's limitations. However, it's the best we've got for now, and still an improvement on Pi's default behaviour on the web.

Thank you for your continued feedback. Feel free to reach out with any issues you encounter.

Warm regards and happy conversing with Pi! :pie:

[1.1.2] - 2023-07-28

Changed

• Auto-submission of transcriptions is now enabled by default. After speaking your message, it will automatically be sent to Pi without needing to press 'Enter'. This feature can be toggled on or off by double-clicking the talk button.

Fixed

• Improved the streaming of transcribed text to be smoother and faster, resolving an issue where the streaming could slow down intermittently.

Enhancements

• Added a more detailed tooltip to the talk button to make the usage of the auto-submit feature more clear.

Version 1.1.0 - Release date: July 27, 2023

Added

• Optional auto-submission feature: This new feature can be enabled by double-clicking the talk button. When active, your transcribed text will automatically be sent without needing to press 'Enter' after each message. A green border around the talk button will indicate when auto-submission is active.

Changed

• Updated the function that handles the transcription response to improve maintainability and readability, in preparation for future enhancements.
• Refactored the functions setNativeValue and simulateTyping to handle the auto-submission feature.
• Updated the event listeners for the talk button to handle the double-click event for activating and deactivating the auto-submission feature.

Fixed

• Improved the method for simulating 'Enter' keypress to work properly with the underlying React application. This was necessary for implementing the auto-submission feature.

Renamed script from saypi.js to saypi.user.js for improved installation experience.