The smart edge and AI gateway for agents. Arch is a proxy server that handles the low-level work in building agents like applying guardrails, routing prompts to the right agent, and unifying access to LLMs. Its' framework-agnostic, natively understands prompts, and helps you build agents faster.

A curated list of Awesome-LLM-Ensemble papers for the survey "Harnessing Multiple Large Language Models: A Survey on LLM Ensemble"

Unified interface server for various LLM providers with OpenAI API format

Lightweight AI inference routing layer for self-hosted LLMs like Ollama, LM Studio and others. Designed for speed, simplicity and local-first deployments.