Baike.dev
All toolsOnline toolsAI codingTrendingOpen sourceNewsSubmit
Log in
< Back to tools
L

loom

> 编程语言
Open source

if your name is not Geoffrey Huntley then do not use loom

1.4K stars0 likes0 views
WebsiteGitHub

About

if your name is not Geoffrey Huntley then do not use loom

Loom

[!CAUTION] Loom is a research project. If your name is not Geoffrey Huntley then do not use.

This software is experimental, unstable, and under active development. APIs will change without notice. Features may be incomplete or broken. There is no support, no documentation guarantees, and no warranty of any kind. Use at your own risk.

Overview

Loom is an AI-powered coding agent built in Rust. It provides a REPL interface for interacting with LLM-powered agents that can execute tools to perform file system operations, code analysis, and other development tasks.

The system is designed around three core principles:

  1. Modularity - Clean separation between core abstractions, LLM providers, and tools
  2. Extensibility - Easy addition of new LLM providers and tools via trait implementations
  3. Reliability - Robust error handling with retry mechanisms and structured logging

Architecture

Loom is organized as a Cargo workspace with 30+ crates:

…

Key Components

Component Description
Core Agent State machine for conversation flow and tool orchestration
LLM Proxy Server-side proxy architecture - API keys never leave the server
Tool System Registry and execution framework for agent capabilities
Weaver Remote execution environments via Kubernetes pods
Thread System Conversation persistence with FTS5 search
Analytics PostHog-style product analytics with identity resolution
Auth OAuth, magic links, ABAC authorization
Feature Flags Runtime feature toggles, experiments, and kill switches

Server-Side LLM Proxy

Loom uses a server-side proxy architecture for all LLM interactions:

┌─────────────┐      HTTP       ┌─────────────┐     Provider API    ┌─────────────┐
│  loom-cli   │ ───────────────▶│ loom-server │ ──────────────────▶ │  Anthropic  │
│             │ /proxy/{provider}│             │                     │   OpenAI    │
│ ProxyLlm-   │  /complete      │  LlmService │                     │    etc.     │
│ Client      │  /stream        │             │                     │             │
└─────────────┘ ◀─────────────  └─────────────┘ ◀────────────────── └─────────────┘
                  SSE stream                        SSE stream

API keys are stored server-side only. Clients communicate through the proxy.

Building

With Nix (Preferred)

Uses cargo2nix for reproducible builds with per-crate caching:

bash
# Build CLI
nix build .#loom-cli-c2n

# Build server
nix build .#loom-server-c2n

# Build any crate
nix build .#<crate-name>-c2n

With Cargo (Development)

bash
# Build everything
cargo build --workspace

# Run tests
cargo test --workspace

# Lint
cargo clippy --workspace -- -D warnings

# Format
cargo fmt --all

# Full check
make check

Specifications

Design documentation lives in specs/. See specs/README.md for a complete index organized by category:

  • Core Architecture
  • LLM Integration
  • Configuration & Security
  • Analytics & Experimentation
  • Editor Integration
  • Remote Execution (Weaver)
  • And more...

License

Proprietary. Copyright (c) 2025 Geoffrey Huntley. All rights reserved.

Issues· 0 open

View all issuesOpen on GitHub

No open issues yet, or sync has not completed.

> Tags

Rust

No comments yet. Be the first to share.

> Details

PublishedAug 1, 2026
UpdatedSep 17, 2026
Category编程语言
PricingOpen source

> Related tools

T
TypeScript
JavaScript 的超集,为前端与全栈提供静态类型
P
Python
通用编程语言,广泛用于 Web、数据与 AI
G
Go
Google 推出的简洁高效系统语言