One config. Every AI harness. Always in sync.
Keep your Claude Code, Copilot, Codex, Cursor, and 7 more AI harnesses synchronized from a single source of truth. Session state bus, per-worktree propagation, and change audit trail. Free tier available.
Most teams using AI have the same config in a dozen different places.
Claude Code, Copilot, Cursor, Windsurf, Aider, Codex, Gemini Code Assist — each with its own config file, each drifting out of sync as rules evolve.
One rule change means 11+ files to update by hand. In practice, most are forgotten. Your agents operate on different rules — even on the same codebase.
Governance policies, model assignments, and context rules live in scattered files. Auditing what each agent actually knows is impossible without a sync layer.
One source of truth. Every harness stays current.
Keep your local AI harnesses in sync from a single config file.
Full harness governance for teams with multiple agents and worktrees.
One config file. One command. All harnesses updated.
Write your governance rules, model assignments, and context instructions in a single YAML config. This is the source of truth for all your AI tools.
Run agent-sync from your project root. It reads your config and writes the correct format to each harness — CLAUDE.md, .cursorrules, .github/copilot-instructions.md, and more.
When rules change, update your config and re-run. Every harness picks up the change in seconds. No manual file hunting, no version drift.
Works with the tools your team already uses.
Free Sync CLI — 11 harness adapters, email-gated download
Pro Suite with worktree fan-out + drift detection dashboard
Enterprise: custom adapters, SSO, team-wide sync policies
Alpha members get the free CLI first and direct input on which harnesses we support next. If you use multiple AI coding tools, this was made for you.
