Skip to content

Where AI
work compounds.

Origin carries sessions, decisions, lessons, project context, and wiki pages across chats, tools, projects, and time.

Get startedView on GitHub
Claude Code pluginClaude Code memoryMCP serverLocal daemon

The problem

Every new AI session starts cold.

The work happened, but the context did not survive. Decisions, fixes, and project instincts stay trapped in old chats instead of helping the next agent.

One missing handoff is enough to make the next conversation repeat the last one.

What Origin brings

A handoff loop for AI work.

Origin captures decisions, lessons, and next steps as work happens, then loads the handoff when the next agent starts.

The next conversation starts from the handoff instead of reconstructing the past.

Background distillery

Origin does the memory chores between sessions.

Between runs, the daemon deduplicates repeat facts, links related ideas, resolves loose notes, and preserves where each claim came from.

The next run starts with consolidated context, not transcript residue.

Knowledge pages

The work becomes reusable pages.

Cleaned decisions and lessons become durable pages instead of buried chat logs. They are organized enough for agents to use and concrete enough for humans to read.

Your work stops being transcript history and starts becoming project knowledge.

Hybrid storage

The database is the index. Markdown is the record.

Origin keeps durable artifacts as Markdown pages, handoffs, and source trails you can open, edit, diff, and move. The local database indexes them for agents.

Agents search the index. You control the files.

Hybrid retrieval, measured

96% fewer tokens. Better answers.

Hybrid retrieval finds the right local context without replaying chat history.

StrategyTokens / queryFinds right context
Full replay (no retrieval)4,505Buried in noise
Basic vector search16877%
Origin (hybrid retrieval)16893.6%

Recall@5 on LongMemEval (oracle, 500 Q). LoCoMo Recall@5 is 70.0%, ~168 tokens per recall query. Run the harness yourself.

FAQ

Common questions.

What is Origin?+
Origin gives AI work one local home. It carries sessions, decisions, lessons, project context, and distilled wiki pages across chats, tools, projects, and time. One daemon, ~30 MCP tools, Markdown projection on disk you can read and edit.
How is Origin different from built-in AI memory?+
Built-in memory stores what the AI decided was important. You usually cannot trace it, correct it, or use it from another tool. Origin keeps memory local, visible, editable, and traceable. Every write is a git commit in ~/.origin/.git/, and every distilled page cites the source memory IDs that produced it.
What retrieval quality does Origin reach?+
Hybrid retrieval combines vector search (BGE-Base-EN-v1.5-Q, 768-dim), FTS5, reciprocal-rank fusion, and knowledge-graph context. Recall@5 is 93.6% on LongMemEval (oracle, 500 questions) and 70.0% on LoCoMo, at roughly 168 tokens per recall query. The eval harness ships in the repo at crates/origin-core/src/eval/.
Is my data private?+
Yes. Origin runs on your machine and stores its database locally. No cloud sync or telemetry by default. Local memory setup works without a model or API key. Distill cycles can use an on-device model or an Anthropic key if you opt in.
Is Origin just another memory MCP?+
No. The MCP server is the connector. Origin also includes the local daemon, background distill cycles, libSQL store with DiskANN vectors, FTS5 + knowledge graph, mandatory provenance, real git versioning, and Markdown export paths.
What AI tools work with Origin?+
Claude Code has a marketplace plugin. MCP-compatible clients such as Cursor, Codex, Claude Desktop, VS Code, Gemini CLI, and others connect through Origin's MCP server.
Is Origin a replacement for Notion or Obsidian?+
No. Origin is not a notes app or a writing tool. It captures and refines what you learn from AI conversations. The Markdown projection under ~/.origin/ can be symlinked into Obsidian if you want to read it there.
How do I set it up?+
In Claude Code, run /plugin marketplace add 7xuanlu/origin, then /plugin install origin@7xuanlu, then /init. For other MCP clients, run npx -y @7xuanlu/origin setup first, then ~/.origin/bin/origin mcp add codex, cursor, claude-desktop, vscode, or gemini.
Does Origin work on Windows or Linux?+
Yes. The daemon builds and runs on macOS (arm64, x64), Linux (x86_64, aarch64; glibc), and Windows (x86_64). Service registration uses launchd on macOS, systemd-user on Linux, and Task Scheduler (schtasks) on Windows.
Can I keep work and personal memory separate?+
Yes. Memories, pages, and recalls belong to a space (for example, work, personal, or client-X). Set the active space per shell with ORIGIN_SPACE, or declare them in ~/.origin/spaces.toml. The auto-detector also picks a space from the current repo or workspace.
Is Origin free?+
Yes. Origin is open-source. The local runtime, CLI, MCP server, and Claude Code plugin files in the Origin repo are Apache-2.0.

Open source

Open where it matters.

The local runtime, CLI, MCP server, and Claude Code plugin are Apache-2.0.

Get startedView on GitHub

Get release updates.