Turnbook
Local long-term memory that keeps agent conversations word for word on the PC.
Tested on Linux
Use it
Linux x86_64
Install turnbook via uv tool
uv tool install turnbook
Where the README and reality differ
- README: `--version` works to confirm install
What worked: use `turnbook --help` (bare `turnbook --version` exits 2: subcommand required)
Check that it works
turnbook --help 2>&1The output should contain “turnbook model [gemma|multilingual]”.
First thing to try
turnbook model 2>&1Run turnbook model to get started.
What it does
- Keep every conversation word for wordNot tested
- Share one memory across AI agentsNot tested
- Look up relevant past memory for each messageNot tested
How far pickfit got
Linux x86_64
- Installs
- Starts
- 0 of 3 features work
- Works offline
macOS and Windows: the project says it runs there; pickfit hasn’t tried it.
Good for
- recall where a conversation left off yesterday
- recall which database was picked for a service and why
- recall what changed in auth middleware last week
An alternative to Mem0, Mastra and Supermemory.
How pickfit tested it
Each run starts from a clean machine and follows the steps above. How pickfit verifies
Linux x86_64
Its main function works
- Replayed from scratch
- Works offline
- Checked on 30 Oct 2026
- Version 0.10.4
Differences the test machine hit
- README: turnbook Python API importable after install
In the test: use the tool env python (e.g. /home/sbx/.local/share/uv/tools/turnbook/bin/python) for `import graph_mind`; system python3 cannot see the isolated uv tool env - README: `turnbook-view` opens the memory
In the test: ensure a memory DB exists at ~/.graph-mind before running viewer, else `sqlite3.OperationalError: unable to open database file`