Feature Ideas

Should be able to use the '@' sign to set a reminder back to that object

OpenAI's Whisper model does a great job at taking voice recordings and transcribing them perfectly. This can be great for transcribing meetings or voice notes for quick capture. It would even be cool to take voice recordings from Telegram voice notes or WhatsApp.

First, a thank you: I really like how Capacities is developing, and I appreciate how quickly the product is improving. Thanks for the many improvements this year. I'm a committed Believer subscriber, upgraded to Believer+, and I'm going to stay. Background I want to test the AI features in Capacities and find out where they complement my workflow. In my daily work I use Claude, including Claude's MCP connector for Capacities. That works very well. With the AI chat inside Capacities itself, however, I repeatedly run into problems. Problem 1: The chat works but returns no answer The AI shows activity ("Explored content", "Read notes", "Explored connections", then "Thought") but ends without giving an answer. When I follow up with "?" or a hint, the next run sometimes delivers the answer and sometimes doesn't. This can happen several times in a row within the same chat. Problem 2: Chat stays stuck on "Preparing content" In one of my chats, the status has been stuck on "Preparing content" for about 24 hours. The circular indicator in the input field keeps spinning and never completes. I can't cancel the generation, and I can't send a new question either. The chat is effectively unusable. Possible connection: model selection I always turn "Auto" off and choose the model myself. I want to use either Claude or models hosted in the EU. Since "Auto" doesn't allow me to exclude models, it isn't an option for me. I don't know whether the errors are related to selecting the model manually. I haven't specifically tested the problem in chats using Auto mode. Expected behavior Every request ends in either an answer or a clear error message. A running generation can be cancelled at any time. A stuck chat doesn't permanently block the input field. Suggestion A timeout with an error message and a working cancel button would already solve a lot. It would also be helpful to be able to exclude individual models or providers in Auto mode, for example "EU-hosted only" or "Claude only". Follow-up: AI chat not responding across models (Capacities v1.71.12, desktop) Here are the details: Platform: Desktop app (Electron), Capacities v1.71.12, Linux x86_64, Chrome 148.0.7778.280 Model/provider: Claude Fable 5.1 (Anthropic). Time of the error: approx. 28 Sept 2026, 21:30 (Europe/Zurich). Diagnostics were generated on 29 Sept 2026, 21:49. New chat: The problem also occurs in a new chat, again on 29 Sept 2026, 22:00. Other models: Claude Sonnet 5 shows the same behaviour, both in the original chat and in the new chat. With Mistral Large, no response arrives either, but an error message appears: "Something went wrong: An error occurred." Browser extensions: not applicable, I use the desktop app Behaviour: With Claude models, the chat shows "Thought" and then no response arrives, without an error message. With Mistral Large, the error message above appears. The original request, creating one planning page and five topic pages in a single message, also produced no pages. The assistant replied that the processing time had run out before the write command could be submitted. Nothing was changed. Since the problem persists across three models from two providers, in a new chat and over two days, I suspect a service-side issue. Could you check whether there is a known problem with the AI chat, or a limit on my account or space?

Add comments and link entities from media highlight text in PDFs and link content add comments to Audio and Video at a certain time add comments to images at specific coordinates Additional features: speech-to-text translation entity and text recognition in images ...

I've uploaded multiple PDFs that all shoe "PDF cannot be displayed" but can be opened on web. Because of this I basically have multiple "Untitled" PDFs that have no metadata I can use to know what I'm referencing.

One of the recent updates brought back some flaky operations in the way outlines work. Was working fine, but one of the last few updates broke it. Create new outline item, make it level 2, enter, tab...should now be NORMAL item under the heading level but it's still level 2, having to grab the mouse, hover, select normal, and continue is awkward and, IMO, broken. This happened a while ago, got fixed, is broken again Should be: if I make an outline item level 2, then new item at same level is also a 2, assume I'm setting up the different sections with heading all of the same level, but when I create a new item and indent, then it becomes normal because the assumption is that now I'm creating the content under that level