Thanks for improving Personal Memory. Bug reports, documentation fixes, UI polish, retrieval improvements, connector templates, and agent-workflow improvements are welcome.
- Preserve the local-first default. Do not make a hosted database, hosted vector store, or remote embedding provider required for normal use.
- Do not commit personal memory data. Keep
memory/,.index/, model caches, connector overrides, tokens, and account-specific configuration out of source. - Use the CLI for memory entries. Do not hand-create or edit files under
memory/entries/; see MEMORY-GUARDRAILS.md. - Keep PRs focused. One behavior change or documentation topic per PR is easier to review.
- Be respectful and follow the Code of Conduct.
Open an issue for large behavior changes, retrieval changes, schema changes, or new connector patterns. Small typo fixes and broken-link fixes can go straight to a pull request.
Good issues describe:
- what problem you hit
- what you expected to happen
- what happened instead
- the command or workflow involved
- whether private data was removed from examples
git clone https://github.com/vladimanaev/personal-memory.git
cd personal-memory
npm install
npm run typecheckUseful commands:
npm run memory -- help
npm run memory -- connectors
npm run index
npm startFor UI changes, run npm start and check the app locally at
http://127.0.0.1:4664.
- The change is focused and explained.
-
npm run typecheckpasses. - README or docs are updated for user-facing changes.
- No personal memory data, private connector overrides,
.index/, model caches, credentials, or screenshots with private data are included. - Changes to capture or recall behavior preserve CLI-based memory access.
- New connector templates avoid personal queries, channels, names, or IDs.
Before opening a PR, run:
git status --short
git diff --cached --statMake sure the staged files are source, docs, or sanitized assets only.