I got tired of explaining my project to every new AI session. So I built LockIn.

Description and use

LockIn stores project context and decisions in a shared notebook, enabling AI tools to retrieve and update this information across devices, reducing redundant explanations and preventing bugs from stale decisions.

Problem

Switching AI coding tools repeatedly required re‑explaining project context, leading to duplicated effort, bugs, and lost decisions.

Outcome

LockIn enables AI tools to share and persist project context across machines, reducing redundant explanations, preventing bugs from stale decisions, and providing a verifiable audit trail of decisions and code changes.

Workflow

• 1. Author records a note or decision in LockIn via MCP.
• 2. The note includes repository, commit, and file references.
• 3. An AI agent retrieves relevant notes using search queries.
• 4. The agent records new findings or handoffs, attaching a context receipt that lists delivered pages and their provenance.
• 5. When a decision is replaced, the replacement is witnessed before being accepted; the graph tracks supersedes relationships.
• 6. LockIn presents the current decision chain and any conflicts to the user for review.

Evidence

I built LockIn: an open‑source notebook that my AI tools can write to and read from.

Source

LockIn connects through MCP, the protocol that lets a compatible AI tool call its save, search and handoff tools.

Source

A note can carry the repository, commit and files it describes.

Source

The graph connects the handoff to its task, the task to the decisions it depends on, and each approved replacement to the page it replaces.

Source