ArchGW 0.2.8 is out 🚀 – unifying repeat "low-level" functionality via an AI-native proxy

I am thrilled about our latest release: Arch 0.2.8. Initially we handled calls made to LLMs - to unify key management, track spending consistently, improve resiliency and improve model choice - but we just added support for an ingress listener (on the same running process) to handle both ingress an egress functionality that is common and repeated in application code today - now managed by an intelligent local proxy (in a framework and language agnostic way) that makes building AI applications faster, safer and more consistently between teams.

What's new in 0.2.8.

• Added support for bi-directional traffic as a first step to support Google's A2A
• Improved Arch-Function-Chat 3B LLM for fast routing and common tool calling scenarios
• Support for LLMs hosted on Groq

Core Features:

• 🚦 Routing. Engineered with purpose-built LLMs for fast (<100ms) agent routing and hand-off
• ⚡ Tools Use: For common agentic scenarios Arch clarifies prompts and makes tools calls
• ⛨ Guardrails: Centrally configure and prevent harmful outcomes and enable safe interactions
• 🔗 Access to LLMs: Centralize access and traffic to LLMs with smart retries
• 🕵 Observability: W3C compatible request tracing and LLM metrics
• 🧱 Built on Envoy: Arch runs alongside app servers as a containerized process, and builds on top of Envoy's proven HTTP management and scalability features to handle ingress and egress traffic related to prompts and LLMs.