One Month: From v0.0.1 to v0.6.25
379 PRs, 2 new team members, and 4 major features - here's what we've been building at Archestra

Written by
Matvey Kukuy, CEO
We've been a bit quiet here on the blog for the last month, and I feel like I owe you all an update. In short, it's been a lot.
- We've merged 379 PRs to Archestra!
- Two people joined the core team
- We've released four major features
- 13 contributors merged code to the upstream
A Month Ago: Security Engine
A month ago Archestra was 100% an AI security engine. At the end of October, we were invited to GitHub Friday to demo our security engine to mitigate "The Lethal Trifecta" vulnerability. First, we tricked the AI into leaking sensitive data. Once we connected N8N via Archestra, it successfully prevented the leak:
๐ Cloud-Native MCP Orchestrator for Kubernetes
We've built a cloud-native MCP orchestrator ready to work at scale for a multi-team enterprise environment. We've solved tons of complexity by storing secrets in Vault, supporting Streamable HTTP, SSE, Stdio, Remote MCP servers, and wiring it up with RBAC...

๐ Chat UI
Archestra is designed to be the centralized AI platform for your company, and we didn't expect end-users to use it directlyโmore as a middleman or gateway. After we realized that Archestra was managing MCP servers, secrets, etc., we added a nice, user-friendly chat UI, and eventually Archestra became a nice place to talk to your data via MCP.
Want to create an agent that speaks to BambooHR, ServiceNow, Jira, and MS Teams, and give it to your manager? That's it.

๐ Internal MCP Registry with Governance
A central place to add "approved" MCP servers. We've put a lot of effort into making this UI friendly. The nice thing is that it supports self-made MCP servers, making it an easy way to deploy your MCP and share it with colleagues, yay!

๐ Cost Monitoring, Token Compressor, Model Optimizer, and Observability
Token spending is a significant topic, and we had to respond accordingly: a full-fledged token observability and optimization suite, including:
- Dynamic model switching
- Token-based compression
- Cost monitoring UI
- OpenTelemetry traces
- Prometheus metrics

A Few Words About the Company
Two outstanding engineers joined us:
- Innokentii, previously a Grafana Mimir and OnCall contributor, finally relocated from Singapore and joined us in London
- Anna joined us in Berlin, transitioning from OSS contributor to a full-time core team member ๐
We visited KubeCon NA Atlanta, and "Wow!" We're so impressed by your warm welcome, especially when you came to our booth and said, "Aha! That's exactly what we've been looking for!"
Thank you for your support! Now it's time to get back to building ๐ช