Sign up to save tools and stay up to date with the latest in AI
bg
bg
1

GitHub - katanemo/archgw: Arch is an intelligent prompt gateway. Engineered with (fast) LLMs for the secure handling, robust observability, and seamless integration of prompts with your APIs - outside business logic. Built by the core contributors of Envoy proxy, on Envoy.

Nov 19, 2024 - github.com
Arch is an intelligent Layer 7 distributed proxy designed to protect, observe, and personalize AI agents with APIs. It is built on Envoy Proxy and handles tasks related to the handling and processing of prompts, including detecting and rejecting jailbreak attempts, intelligently calling "backend" APIs, routing to and offering disaster recovery between upstream LLMs, and managing the observability of prompts and LLM interactions. Arch offers features such as built-in Envoy, function calling for fast Agentic and RAG apps, prompt guard, traffic management, and standards-based observability.

The markdown data provides a guide on how to set up Arch and integrate it into generative AI applications. It includes prerequisites, steps to install Arch, configuring Arch with your application, and using OpenAI Client with Arch as an Egress Gateway. The document also mentions that Arch supports best-in-class observability by supporting open standards and welcomes contributions to improve the platform.

Key takeaways:

  • Arch is an intelligent Layer 7 distributed proxy designed to protect, observe, and personalize AI agents with APIs. It is built on Envoy Proxy and handles tasks related to the handling and processing of prompts.
  • Arch offers core features such as built on Envoy, function calling for fast Agentic and RAG apps, prompt guard, traffic management, and standards-based observability.
  • The function calling LLM (Arch-Function) designed for the agentic and RAG scenarios is hosted free of charge in the US-central region. Pricing for the hosted version of Arch-Function will be around $0.10/M output token.
  • Arch is designed to support best-in class observability by supporting open standards. It also welcomes contributions to improve its features, fix bugs, improve documentation, or create tutorials.
View Full Article

Comments (0)

Be the first to comment!