🗺️ Roadmap
Curiositi is actively developed. This roadmap outlines what has been built and what is planned.
Current Status
Section titled “Current Status”Curiositi currently supports the following types of data:
- Text
- Text-based files like Markdown, HTML, etc.
- Images
- JPEG, PNG, GIF, etc.
Planned
Section titled “Planned”Office Docs
Section titled “Office Docs”Support for Word, Excel, and PowerPoint files.
Web Pages
Section titled “Web Pages”This requires scraping the web page, generating metadata for it, and then generating embeddings for the content and the metadata. Ideally this would use Firecrawl.
This requires transcribing the audio, generating metadata for it, and then generating embeddings for the content and the metadata.
- User Uploaded Videos — Transcribe the audio, generate metadata, and generate embeddings for the content and metadata. Extract frames from the video, pass them into a Vision-capable model, and generate embeddings for them (similar to image embedding).
- YouTube Videos — Download the videos, transcribe the audio, generate metadata, and generate embeddings. Extract frames and generate embeddings using a Vision-capable model.
Browser Extension
Section titled “Browser Extension”This would allow users to save any text they come across on the web.
Mobile App
Section titled “Mobile App”This would function exactly the same as the web app.
Contributing
Section titled “Contributing”Want to contribute or suggest features?