Skip to content

Releases

v1.0.0 — Official Launch of VeloRoute AI Gateway

Choose a tag to compare

@mrgusux mrgusux released this 08 Nov 17:27

⚡ VeloRoute v1.0.0 — Production Release

We are excited to announce the official v1.0.0 release of VeloRoute, the enterprise-grade universal AI Gateway, high-availability router, and Token-Saviour optimization engine.

🌟 Highlights & Key Capabilities

  • Universal Multi-Provider Gateway: Drop-in OpenAI /v1/chat/completions, native Anthropic Claude /v1/messages, Gemini /v1beta/models, and Ollama /api/chat reverse proxy.
  • Token-Saviour Suite: Heuristic AST code skeletonization, Caveman terse mode, and lossless terminal trim cutting token overhead by up to 76%.
  • KV-Cache Prefix Affinity: Intelligent session cache pinning yielding up to 90% prompt caching cost reductions.
  • Sentinel Circuit Breakers: Sliding-window key health tracking, half-open probe recovery, and zero-downtime failover.
  • Sovereign Privacy Shield: Zero-retention memory mode preventing prompts or completions from touching storage.
  • Model Context Protocol (MCP): Built-in Anthropic/OpenAI standard JSON-RPC 2.0 gateway with SSE transport.
  • Modern Next.js 15 Landing Page: Deployed live on Vercel at https://veloroute-tau.vercel.app.
  • Docker & Cloud Native: Multi-stage lightweight container (ghcr.io/mrgusux/veloroute:latest), docker-compose.yml, and enterprise Helm charts (deploy/helm/veloroute).

🚀 1-Command Quickstart

�ash docker compose up -d