What is RunTrim?
RunTrim is the control, verification and recovery layer for AI coding agents. It gives coding agents scope, memory, finish checks, restore points and continuation prompts while keeping source code local by default.
RunTrim is the control, verification and recovery layer for AI coding agents. It gives coding agents scope, memory, finish checks, restore points and continuation prompts while keeping source code local by default.
No. Source code stays local by default. Cloud sync stores run metadata only.
Run metadata such as verdict, changed file paths, project memory summaries, and restore metadata. Not raw source file contents.
RunTrim checks sensitive paths for safety, but does not read or upload env file contents.
Yes. Free runs locally with guarded runs, finish verification, and local restore.
RunTrim works with Claude, Codex, Cursor, ChatGPT, DeepSeek, and other coding agents through copy mode or MCP-compatible workflows.
No. MCP is optional. Copy mode works with any agent UI.
No. RunTrim is the control, verification and recovery layer around your agent.
RunTrim is not an orchestrator that routes model calls for you. It is a guard layer that adds scope contracts, verification and recovery around the agent workflow you already use.
runtrim start runtrim agent "your task" --copy runtrim finish runtrim restore, if recovery is needed
No. Restore is local recovery and does not require another agent run.
Usually yes. By keeping runs scoped and carrying forward project memory, RunTrim often reduces repeated context loading and shortens rework cycles. Results vary by workflow.
BLOCKED means the run needs review and is not trusted yet. Review the changes, approve a scoped change, or restore.
Restore apply is local through the CLI. The dashboard shows restore metadata and recovery guidance.
Restore metadata is the run-linked recovery record shown in dashboard history. File recovery still happens locally.
Free protects one project locally. Pro adds cloud sync and dashboard history for your main project. Builder adds unlimited projects and CI Gate. Team adds governance controls, with some capabilities marked reviewed access or coming soon.
Git tracks version history. RunTrim adds run-level control, verification and local restore for AI coding agent workflows before and after changes are made.
Free is designed for local-first guarded runs. Pro unlocks cloud run history, memory sync, and restore metadata sync.
Upgrade when you need multi-project protection, advanced recovery history, and CI merge gating.
Team is reviewed access. Some governance features are live or rolling out, and unfinished capabilities should be labeled coming soon.
Yes, with runtrim ci check when configured. BLOCKED runs should fail the check.
No. CI and policy depth are plan-gated. Builder unlocks CI Gate, and Team expands governance.
Only show implemented controls as live. Label shared logs, approvals, and policy surfaces accurately if still rolling out.
Start local with the CLI, then unlock cloud history and governance when your workflow needs it.