loom
Loop engineering for agentic software delivery.
Loom is an MCP server that implements loop engineering for agentic software delivery, helping coding agents complete larger tasks without losing state. It maintains project context across sessions by storing requirements, task progress, test results, and repair notes in a local `.loom/` directory. The server works with Claude Code, Codex, OpenCode, and other MCP-capable coding agents to provide a structured delivery workflow—plan, build, test, fix, preview, and handoff—backed by persistent state and targeted context routing. Built with a Rust MCP runtime and Python algorithm layer, it transforms one-shot coding sessions into resumable delivery loops with verification, review, and deployment support.
Key Features
Use Cases
- 01Completing feature work that spans multiple coding sessions without losing context or requirements
- 02Managing complex software delivery tasks with verification, review, and repair workflows
- 03Resuming interrupted agent sessions from saved delivery state instead of restarting from scratch
- 04Generating local Docker Compose deployment previews with proper service topology and configuration
- 05Integrating domain knowledge and product documentation into agent clarification workflows
- 06Coordinating multi-step tasks across different coding agents using shared delivery state
Related MCP Servers
View moremarkitdown
Python tool for converting files and office documents to Markdown.
firecrawl
The context API to search, scrape, and interact with the web at scale. 🔥
prompts.chat
f.k.a. Awesome ChatGPT Prompts. Share, discover, and collect prompts from the community. Free and open source — self-host for your organization with complete privacy.
langflow
Langflow is a powerful tool for building and deploying AI-powered agents and workflows.
loom — FAQ
What is the Loom MCP server?+
Loom is an MCP server that provides loop engineering for coding agents, maintaining persistent delivery state across sessions. It helps agents complete larger software tasks by managing requirements, task progress, testing, review, and deployment through a structured workflow that survives interruptions and context loss.
How do I install the Loom MCP server?+
Run the installer script with your agent choice: `curl -fsSL https://github.com/valkor-ai/loom/releases/latest/download/install.sh | bash -s -- --agent codex` (or claude-code, opencode, or all). The installer detects your platform, downloads the release, installs the Rust MCP server, registers it with your agent, and runs verification. On Windows, use the PowerShell installer script instead.
Which AI clients work with Loom?+
Loom currently supports Codex, Claude Code, and OpenCode through their MCP integrations. It works with any MCP-capable coding agent that can register and communicate with MCP servers.
Does Loom require API keys or external services?+
No, Loom runs entirely locally as an MCP server. It requires Docker if you want to use the deployment preview features, but otherwise operates on local project state without external API dependencies.
Is the Loom MCP server free?+
Yes, Loom is open source software released under the Apache License 2.0 and is free to use.
What happens if my coding session is interrupted?+
Loom stores all delivery state locally in the `.loom/` directory, including requirements, task plans, results, review notes, and repair requests. When you reopen your agent and run the continue command, Loom resumes from exactly where the previous session stopped without rebuilding context.
How do I install loom?+
Open the source repository on GitHub and follow its README. loom is a mcp server — MCP Agents Market links you directly to the official repo.
Is loom free?+
loom is an open-source project hosted on GitHub. Check the repository for its license and any usage requirements.