Canari shows a local-first agent design that may cut memory token use
Canari is a AI agent desktop app for Windows. It can run with Ollama models on the user’s own computer, , or a mix of both. It can read, write, edit, list, and delete files, run persistent Python, and search a project.
It also supports web search, page fetching, screen screenshots with vision, Telegram messages, opening , and PDF export. Its default uses an , where it can save, add to, search, and delete notes with tags and . Memory search combines BM25 with optional , so the agent can find the right note without putting the full history into the context on every turn.
Dream Mode reviews work after the user has been away, creates ideas, saves them to Obsidian, and sends a Telegram alert. Daily Resume summarizes the previous day on the first launch of the day, and Weekly Scan suggests merging near-duplicate notes every seven days with user approval.
Key points
- Canari is a Windows desktop app for building and running a AI agent.
- It can use local Ollama models, , or both together.
- Obsidian is used as the default memory store for notes.
- BM25 and help retrieve specific memories without sending the full history as context.
- It includes daily summaries, duplicate-note review, and an idle reflection mode called Dream Mode.