The article provides a quickstart guide for using Arch Gateway to build AI agents, such as a currency exchange agent, and demonstrates its use as an LLM router. It includes steps for setting up Arch with Docker, creating configuration files, and interacting with the gateway using curl commands or the OpenAI Python client. Arch supports observability through integration with tools like Signoz and encourages community contributions to improve its features and documentation.
Key takeaways:
- Arch is an intelligent proxy designed to enhance agentic applications by connecting existing APIs, providing features like intent-based prompt routing, prompt guard, LLM routing, and observability.
- Arch Gateway is built on Envoy Proxy, offering secure handling, intelligent routing, and robust observability for prompt-based tasks, with purpose-built LLMs to manage prompts effectively.
- The platform supports building AI agents with a simple configuration process, allowing integration with APIs like currency exchange services, and offers a free hosted version of its function-calling LLM in the US-central region.
- Arch provides observability features using open standards, enabling request tracing, metrics monitoring, and integration with tools like Signoz for enhanced AI application performance.