Loom

Loom is an open-source AI agent harness and developer orchestration platform built primarily in Rust by Geoffrey Huntley. It provides a centralized server and CLI client architecture designed to manage, coordinate, and govern AI-assisted engineering workflows across development environments. In day-to-day use, developers interact with Loom through its cross-platform terminal binary, which connects to a Loom backend server. The system acts as a managed runtime and LLM proxy, orchestrating agent workflows, managing conversation threads, and routing sync and streaming requests to providers like Anthropic. It features dedicated authentication mechanisms, including an Anthropic OAuth integration flow, device code flows, and API keys. Loom distinguishes itself from basic standalone CLI coding agents by focusing heavily on infrastructure governance, proxying, and security. It includes native audit logging for all proxy endpoints, automatic secret redaction (`loom-redact`) across actions and user data, enterprise identity providers (Okta, GitHub, Google), webhook lifecycle hooks, and self-updating client binaries across Linux, macOS, and Windows. The project is open-source under active development on GitHub, positioned as an enterprise-grade agent orchestration substrate for engineering teams rather than merely an unmanaged prompt runner.

Tags: cli, developer tools, experiments, OpenSource

Visit Loom

Key features

  • Centralized LLM Proxy Engine — Brokers synchronous and streaming model requests with complete request tracking and inspection. - Automatic Secret Redaction — Integrates loom-redact into audit logging pipelines to automatically sanitize keys, tokens, and credentials. - Cross-Platform CLI Client — Distributed for macOS, Linux, and Windows with embedded build metadata and self-updating (loom update) capabilities. - Extensible Identity & Auth — Supports Anthropic OAuth two-step browser flows, GitHub, Google, Okta, magic links, and device codes. - Agent Workflow Automation — Runs automated agent routines and thread-based sessions defined in repository workflow directories. - Comprehensive Audit Trail — Captures detailed audit logs for logins, LLM proxy usage, API key access, repository mirroring, and webhooks.

Use cases

Use Case: Centralized Enterprise Model Proxy Route developer terminal interactions and automated agents through a unified Loom instance to enforce credential control, track token usage, and automatically strip credentials from log streams.

Use Case: Automated Repository Maintenance Workflows Configure and run persistent agent automation routines (such as automated updates and dependency maintenance) directly orchestrated against repository targets.

Use Case: Governed Multi-User Team Access Equip engineering teams with standardized CLI agent access using centralized corporate SSO and OAuth credentials rather than distributing raw LLM vendor API tokens.

Who it is for

Loom is built for platform engineers, DevOps leads, and software teams requiring centralized governance, security auditing, and proxy controls around developer AI agent usage. It is not intended for individual developers looking for a minimalist, single-file CLI wrapper that requires no server infrastructure.

Install guide

Installation Build the Loom CLI client and server from source using Cargo: bash git clone https://github.com/ghuntley/loom.git cd loom cargo build --release When connected to an active Loom server deployment, update the binary directly: bash loom update

First Run Verify client installation and configure authentication against your Loom server: bash loom version loom auth login

Example Session Trigger an agent workflow managed by Loom: bash loom workflow run .agents/workflows/nixos-auto-update.yml

Part of CLI Coding Agents

Categories: AI Coding Agents, Developer Tools, Experiments

Related CLI agents

  • Catnip — Catnip is an agent harness and orchestrator project under the Weights & Biases (wandb) organization designed to manage, execute, and…
  • g3 — g3 is an open-source, Rust-based command-line coding agent framework developed by Dhanji Prasanna. It is designed to orchestrate LLM-driven software…
  • Unship — Unship is an open-source agent harness and plugin integration framework developed by mbenhard. It is designed to package and orchestrate…
  • TeDDy — TeDDy (developed by atte500) is an open-source test-driven execution harness and orchestrator written in Python. It is designed to manage…
  • wreckit — wreckit is an open-source CLI agent harness and orchestration framework created by Mike Hostetler. Built using TypeScript and designed around…
  • Ouroboros — Ouroboros is an open-source autonomous CLI coding agent developed by razzant. Built primarily in Python, the agent is engineered to…
  • AI Tools
  • Categories
  • Industries
  • CLI Coding Agents
  • MCP Servers
  • MCP Categories