OpenSelf
User-owned local context vault that stores memories and compiles bounded context for AI agents.
Tested on Linux
Use it
Linux x86_64
You need first
- Runtimes:
Node.js 22.13 or newer
Install openself via npm
npm install -g openself
Where the README and reality differ
- README: openself init --data-dir ~/.openself and openself doctor
What worked: no init needed; `openself memory add --data-dir /tmp/osvault` auto-creates vault; `openself init` and `openself doctor` do not exist in 1.0.1 - README: 1.1.0 stable, openself context --explain, openself demo --data-dir ./demo-data
What worked: npm latest is 1.0.1 (only 0.7.0, 1.0.0-rc.2, 1.0.0, 1.0.1 published); `openself context --explain` and `openself demo` do not exist (unknown command)
Check that it works
openself --versionThe output should contain “1.0.1”.
First thing to try
rm -rf /tmp/osvault && openself memory add --type decision --scope project/openself --content "Use SQLite as the local source of truth" --source docs/architecture.md --tags architecture,database --data-dir /tmp/osvault && openself memory search --query "Which database did we choose?" --scope projectRun openself memory add with a type, scope and content to store your first memory.
What it does
- Store decisions, preferences and factsTested
How to use it
- Add a memory:
openself memory add --type decision --scope <scope> --content "<your text>" --source <file> --tags <tags> --data-dir <dir> - Search it:
openself memory search --query "<your question>" --scope <scope> --data-dir <dir> - See vault totals:
openself memory stats --data-dir <dir>
- Add a memory:
- Import documents and chat exportsTested
How to use it
- Import a file:
openself memory import --file <file> --scope <scope> --data-dir <dir> - Search the imported memories:
openself memory search --query "<your question>" --scope <scope> --data-dir <dir>
- Import a file:
- Build size-limited context packages with explanationsTest failed
How far pickfit got
Linux x86_64
- Installs
- Starts
- 2 of 3 features work
- Works offline
macOS and Windows: the project says it runs there; pickfit hasn’t tried it.
Good for
- store a durable decision as a typed memory
- search stored memories by query and scope
- preview exact context block an agent would receive with receipt
- import documents and WhatsApp or Telegram chat exports
- review and approve agent-proposed memories
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 1.0.1