Jotit – command-line notes with AI search and summaries (github.com)

Jotit is a new open-source command‑line note tool that combines instant note capture with AI-powered semantic search and automated summaries to remove the friction of naming and organizing files. Instead of choosing folders and filenames, you jot text from the shell and later query it with natural‑language search, get periodic AI summaries (e.g., last 5 days, week, month), or scope searches by tag/date. It’s meant to speed capture and make retrieval smarter by using embeddings for concept matching (not just keywords) and summarization to surface trends in your notes. Under the hood Jotit stores notes in a local SQLite database that supports vector search, runs embedding generation as background jobs (on demand or via a worker), and can be configured via ~/.config/jotit/jotit.toml or environment variables. It requires Python 3.10+, is installed from the GitHub repo (not on PyPI), and uses an Anthropic API key for AI summaries/QA. Defaults include the all‑MiniLM‑L6‑v2 embedding model and a Claude Sonnet summarization model; search supports a tunable similarity threshold and hybrid (semantic+keyword) mode. CLI commands cover add/list/search/summary/worker/edit/delete/import and developer tooling (just, pytest, mypy); the project is MIT‑licensed.