Project identity
One local identity follows a project across agents, sessions, and Git worktrees.
Local context for coding agents
Agents and sessions change. Continuity keeps the project's memory, evidence, and next steps in one local place, ready for whoever picks up the work.
Open source. Runs locally. Works across agents.
Why Continuity
Every coding agent brings its own tools. Continuity gives them a shared, project-bound record of what has been learned and what needs to happen next.
One local identity follows a project across agents, sessions, and Git worktrees.
Keep source-backed facts and attributed lessons with their trust level and history.
Pass on the goal, completed work, remaining steps, decisions, and risks.
See which current sources support context, and why an item was selected.
How it works
One project record carries the context forward. Select a step to see the commands and the information the next session can use.
Illustrative flow. Commands run in your own project.
$ continuity init
$ continuity sync
Project identity stays bound to this directory. Sources are checked before context is reused.
Your files and Git remain the source of truth.
$ continuity handoff create --file handoff.json
The record keeps its origin and capture time.
$ continuity handoff latest
Recheck files and tests before acting on earlier claims.
Built for real projects
No hosted account or model key is required for the default experience. Git and your files remain authoritative.
The trusted host fixes project scope. Context from another project cannot be retrieved implicitly.
Memory keeps provenance and revision history. Conflicting claims are quarantined for review.
Start with a project you already have.