Changelog (Nightly)
🚧 Important Notice
Please do not share the nightly version.
If you encounter issues and bugs related to the nightly build, please don’t share them publicly in the community, but report them in the #nightly-tester channel.
All other changes are on our What’s New page.
v1.55.11 (2025-11-12)
New AI infrastructure and additional model providers
This release introduces a major update to our AI infrastructure, enabling a highly-anticipated feature: the ability to add and manage API keys from other top-tier LLM providers! Now you can leverage the models you prefer directly within Capacities. In particular, users will now be able to select and add API keys from: OpenAI, Gemini, Anthropic, Mistral and xAI.
How to Connect Your Key:
- Go to Settings > AI Assistant > Custom AI Provider.
- Click “+ Add new AI key”.
Once added, you can tailor your key’s usage with new management options:
- Key Name (Alias): Set a key name (e.g., “Work,” “Personal”) for easy identification.
- Available Spaces: Restrict the key’s use to a specific space in your account.
- Max Budget: Set a monthly spending limit (in USD) for usage through Capacities (unlimited by default).
Note: You can still opt to use only your private keys at the bottom of the AI Assistant settings page.
The AI Assistant model dropdown will automatically show the models available for the keys you have enabled in that specific space. If you have multiple keys from the same provider, the custom key name will be displayed next to the model.
We’d love to hear your thoughts!
- Is the key integration process intuitive?
- Are you missing any available models from your provider of choice?
- Are there any other LLM providers you are missing?
Please also let us know if you encounter any issues.
Kindle Integration
We are launching an import/export integration for Kindle Highlights soon. You can already test it. More info here.
We’d love to hear your feedback:
- Does everything work?
- Is something confusing?
- Do you like the integration?
Web Highlights 2.0
We also updated the Web Highlights integration. It works exactly as it did before but now also support the highlights you know from the Readwise integration.
Please let us know if you encounter any issues.