# archestra

> archestra is an open-source AI tool for security. Open-source gateway adding identity, policy and audit between AI agents and your systems. It is available on self-hosted and API.

- **Kind:** Tool
- **Pricing:** Open source
- **Platforms:** self-hosted, API
- **Website:** https://archestra.ai
- **Repository:** https://github.com/archestra-ai/archestra
- **GitHub stars:** 4,444
- **Categories:** security, AI models and LLM tooling, developer tools
- **Listing updated:** 2026-11-14

## About

Archestra is built for platform teams rolling AI out inside a company. It sits between AI clients such as Claude Code, Codex, Cursor, Copilot CLI or n8n and the models, MCP servers and business systems they reach, with one registry, gateway and audit trail for MCP and one endpoint for approved LLMs.

Each tool call is tied to a user identity through Entra ID, Okta, OIDC or SAML. Policies run before execution to control tool access, approvals, isolation and network egress. The LLM proxy keeps API keys server-side, sets spend limits and routes across providers including OpenAI, Anthropic, Azure OpenAI, Vertex AI and Bedrock. Telemetry exports via OpenTelemetry and Prometheus.

It deploys locally, from source or on Kubernetes with Helm and Postgres, under AGPL-3.0.

## Features

- **MCP gateway:** One registry, identity layer and audit trail for MCP servers.
- **LLM proxy:** Single endpoint with provider routing, server-side keys and spend limits.
- **Pre-execution policies:** Controls tool access, approvals, isolation and egress.
- **Identity-aware calls:** Entra ID, Okta, OIDC or SAML tie actions to users.
- **Observability:** Traces requests and tool calls, exports OpenTelemetry and Prometheus.

## Good for

- Rolling out MCP servers with central access control
- Giving teams one approved LLM endpoint
- Auditing who ran which agent tool call
- Tracking and capping AI spend

## FAQ

### What does it cost?

It is AGPL-3.0 and free to self-host. SSO/OIDC, RBAC and white-labelling need an enterprise licence, free for teams under 30 users.

### Where does it run?

Locally, from source or on Kubernetes.

### Which providers work?

OpenAI, Anthropic, Azure OpenAI, Vertex AI, Bedrock and self-hosted models.

### Do I need sales?

The FAQ says no.

### Is archestra free?

Yes. archestra is open source, so you can use and self-host it for free.

### What are the best alternatives to archestra?

The closest alternatives to archestra are ECC, firecrawl, ponytail, graphify and orca.

## Alternatives

- [ECC](https://ai-powered.com/tools/ecc): ECC is a freemium AI tool for ai models and llm tooling.
- [firecrawl](https://ai-powered.com/tools/firecrawl-2): firecrawl is a freemium AI tool for web search and scraping.
- [ponytail](https://ai-powered.com/tools/ponytail): ponytail is an open-source AI tool for ai models and llm tooling.
- [graphify](https://ai-powered.com/tools/graphify-2): graphify is a freemium AI tool for knowledge and memory.
- [orca](https://ai-powered.com/tools/orca): orca is an open-source AI tool for coding.
- [archify](https://ai-powered.com/tools/archify): archify is an open-source AI tool for design and media.
- [ragflow](https://ai-powered.com/tools/ragflow): ragflow is a freemium AI tool for knowledge and memory.
- [Agent-Reach](https://ai-powered.com/tools/agent-reach): Agent-Reach is an open-source AI tool for ai models and llm tooling.

---
Source: https://ai-powered.com/tools/archestra
