Archestra

Archestra is an open-source, self-hosted enterprise MCP (Model Context Protocol) platform that enables organizations to securely deploy AI agents at scale. The platform provides agentic chat for non-technical users with SSO authentication through Entra ID, Okta, or any OIDC provider, with every tool call running under the user's own identity. Apps and skills render real UI inside conversations including dashboards, boards, and forms, while projects provide shared context across chats, files, and scheduled tasks. the Agent Runtime ships agents to server-side sandboxed containers with per-user execution and cost limits, the MCP Orchestrator runs MCP servers as containers in Kubernetes with security review and approval workflows, RAG connects Confluence, Jira, and internal docs while preserving source system permissions, LLM and MCP proxies manage secrets with virtual keys and and static guardrails prevent data exfiltration by tainting conversations when sensitive data is returned. Every token, tool call, and dollar is attributed to a person, team, and agent with OpenTelemetry export to Grafana or Splunk. Tested with Fortune 50 companies, Archestra has a open-source community of 4,000 members. Founded in 2025 and headquartered in London, United Kingdom, the company has raised approximately $13.5 million in seed funding from investors including Concept Ventures, Zero Prime Ventures, Celero Ventures, RTP Global, 20VC Fund, Visible Ventures, and Tenacity Capital.

AI Infrastructure > Enabling Technologies > Software > ML Orchestration > AI Agent Frameworks

Company Wise > Cisco