Loom
LLM agent framework with natural language agent creation
Loom is a Go framework for building autonomous LLM agent threads with natural language agent creation, pattern-guided learning, and multi-agent orchestration. It includes: - Natural language agent creation via the weaver meta-agent - Multi-agent orchestration with 6 workflow patterns - 90+ reusable patterns across 16 domains - 8 LLM provider integrations (Anthropic, Bedrock, OpenAI, etc.) - Multi-modal support (vision, document parsing) - Built-in judge evaluation system with DSPy integration - Self-improving agents with pattern proposal capabilities - Real-time streaming and complete observability
winget install --id Teradata.Loom --exact --source winget Latest 1.0.2
| Architecture | Scope | Download | SHA256 |
|---|---|---|---|
x64 | — | Download | 33D2D3C5F83D0EB04B1318225B4E1C891405D554BC86D4168E4B5DD53E30FBC3 |
Details
- Homepage
- https://github.com/teradata-labs/loom
- License
- Apache-2.0
- Publisher
- Teradata
- Support
- https://github.com/teradata-labs/loom/issues
- Moniker
loom