# Gatana

> Gatana is an MCP gateway that offers turn-key support for skills, SSO, and more.

Gatana is an MCP gateway that offers turn-key support for skills, SSO, and more. It sits between your AI agents, such as Claude, ChatGPT, Codex and Cursor, and the tools they call. Every call runs under the user’s own identity, credentials stay out of local machines, and every action is recorded for audit. Gatana Cloud runs in the EU; Enterprise customers self-host on Kubernetes with the official Helm chart. Plans: Free ($0, 1 user, unlimited remote MCP or OpenAPI servers), Pro ($125 per month per organization, unlimited users, local and Runnable Code servers, 20 GiB persistent storage, Tailscale VPN, federated identity with SAML, OIDC and SCIM, SIEM streaming, code mode, AI compression, AI sandboxes in early access), Enterprise (custom, on-premise, dedicated support). The Pro trial runs 14 days with no credit card.

## Product

- [Home](https://www.gatana.ai/): what Gatana does, which agents it works with, pricing and FAQ
- [Pricing](https://www.gatana.ai/pricing): plans, limits and a comparison table
- [Pricing as Markdown](https://www.gatana.ai/pricing.md): the same plans and limits in a machine-readable file
- [Measurements](https://www.gatana.ai/measurements): numbers we measured ourselves, each with the date, the method and the raw data
- [Measurements as Markdown](https://www.gatana.ai/measurements.md): the same figures in a machine-readable file
- [Security](https://www.gatana.ai/security): vulnerability reporting, how customer data is protected, incident notification
- [Support](https://www.gatana.ai/support): how to reach support

## Documentation

- [Docs home](https://docs.gatana.ai/): quick start and reference
- [Docs index for language models](https://docs.gatana.ai/llms.txt): every documentation page with a one-line description
- [All documentation in one file](https://docs.gatana.ai/llms-full.txt)
- [Gateway introduction](https://docs.gatana.ai/gateway-introduction/): how servers, tools and the gateway fit together
- [Adding MCP servers](https://docs.gatana.ai/servers-standard-mcp/): remote HTTP servers and local uvx or npx executables
- [OpenAPI and Swagger servers](https://docs.gatana.ai/servers-openapi-swagger/): turn an API specification into tools
- [Runnable Code](https://docs.gatana.ai/servers-runnable-code/): serverless Node.js and Python functions as tools
- [Skills](https://docs.gatana.ai/skills/): create, share and keep skills in sync across agents
- [Profiles](https://docs.gatana.ai/profiles/): credential and access control per user, team or token claim
- [Tool firewall](https://docs.gatana.ai/tool-firewall/): deny or log tool calls by rule on arguments, user, team or token claim
- [Federated identity](https://docs.gatana.ai/federation/): SAML, OIDC and SCIM with Okta, Entra ID and others
- [MCP authorization](https://docs.gatana.ai/mcp-authorization/): how agents authenticate to the gateway
- [Audit logs](https://docs.gatana.ai/audit-logs/): what is recorded for every call
- [SIEM streaming](https://docs.gatana.ai/siem-streaming/): HMAC-signed NDJSON to Splunk, Falcon LogScale, Elastic, Sentinel or any collector
- [On-premise self-hosting](https://docs.gatana.ai/on-premise-self-hosting/): running Gatana on your own Kubernetes cluster
- [Public API](https://docs.gatana.ai/public-api-introduction/): REST API, OpenAPI specification, SDKs

## Quick-Start: Use Gatana as an MCP Gateway

You can add https://gatana.ai/mcp as a remote server to any agent that supports the MCP protocol.
During authorization, user has possibility of creating a new organization. The organization ID
can be used to get the Self-Service Admin web dashboard: https://ORG_ID.gatana.ai/

## Gatana CLI

curl -fsSL https://github.com/gatana-ai/gatana/releases/latest/download/install.sh | sh
gatana config login
gatana get servers

## Guides

Reference pages, one per question, kept current; the date on each page is the date of its last revision.

- [How Much Does an MCP Gateway Cost? List Prices Compared](https://www.gatana.ai/guides/how-much-does-an-mcp-gateway-cost): Vendors price MCP gateways in four ways: open source that you host yourself, a price for each user, a price for each organization, and a price for each tool call. This guide lists the published prices of ten products on 6 October 2026. It calculates the cost for a team of 25 people. It also names the costs that are not on any price list.
- [What Is an MCP Gateway?](https://www.gatana.ai/guides/what-is-an-mcp-gateway): An MCP gateway is one server between your AI agents and the MCP servers, APIs and functions that they call. It handles identity, credentials, access rules and audit for all of them. This guide explains what a gateway does and when a team needs one. It also gives the cost in latency and tokens, and how the options differ.
- [What Is Code Mode for MCP?](https://www.gatana.ai/guides/what-is-code-mode): Code Mode allows an AI agent to work with thousands of tools without consuming any more context than a static fixed context cost of two (2) tools: search and execute.
- [How to Add SSO, SAML and SCIM to MCP Servers](https://www.gatana.ai/guides/how-to-add-sso-saml-scim-to-mcp-servers): MCP servers do not speak SAML or SCIM. A gateway in front of them does. This guide shows the four parts of the setup and the steps in Gatana. It also shows what nine gateway vendors publish about SAML, OIDC and SCIM support as of October 2026.
- [How to Get Audit Logs of Every MCP Tool Call Into Your SIEM](https://www.gatana.ai/guides/mcp-audit-logs-to-siem): Agents write logs on laptops. MCP servers see tokens, and not people. Only a gateway can make one record for each tool call with the person behind it. This guide shows what a record must contain. It also shows how to stream the records to Splunk, Falcon LogScale, Elastic or Sentinel, and what the delivery guarantees are.
- [MCP Gateway With EU Data Residency: Where Gatana Keeps Your Data](https://www.gatana.ai/guides/mcp-gateway-eu-data-residency): An MCP gateway holds three kinds of sensitive data: stored credentials, the content of tool calls, and the audit trail. This guide shows where Gatana Cloud processes each kind. It also shows which sub-processors touch the data, what crosses the EU border, and how to keep all of the data inside your own network.
- [How to Run Serverless Functions as MCP Tools](https://www.gatana.ai/guides/run-serverless-functions-as-mcp-tools): Not every system has an MCP server or an OpenAPI specification. A gateway that hosts code lets you write a function and expose it as a tool. This guide shows the three ways to do it. It also shows how Runnable Code servers work in Gatana and the limits to know before you select a method.

## Blog

- [Code Mode vs. Direct Tool Calls: We Measured](https://www.gatana.ai/blog/code-mode-vs-direct-tool-calls-benchmark): One model, three data tasks, five runs each: tools called directly, then through Gatana's code mode. On 10,000 orders, direct calls were right 2 times out of 5 at 1.3 million prompt tokens per run. Code mode was right 5 of 5 at 30,000 tokens, in 12 seconds instead of 10 minutes.
- [How Much Latency Does an MCP Gateway Add? 6.6 Milliseconds in Our Latest Test.](https://www.gatana.ai/blog/how-much-latency-does-an-mcp-gateway-add): Our latest test measured 6.6 ms of added latency per tool call on Hetzner with an AMD Ryzen™ 5 3600, with the gateway and the upstream server on MCP revision 2026-07-28. The same build adds 9.6 ms against a 2025-era server. We made 21,900 timed calls across three repetitions. Here are the setup, exact figures, and raw data.
- [We Converted 300 Public OpenAPI Specs to MCP Tools. Here Is What Broke.](https://www.gatana.ai/blog/openapi-to-mcp-conversion-300-public-specs): A measured answer to how well OpenAPI converts to MCP: 262 of 300 public specs converted on the first run, 299 after we fixed four defects in our own converter, and the generated schemas shrank eightfold. Method, numbers and the per-spec data are included.
- [Skills: Teach Every Agent How Your Organization Works](https://www.gatana.ai/blog/skills-teach-every-agent-how-your-organization-works): A skill is a Markdown document that tells an agent how a job is done in your organization. Gatana now stores your skills in one place, installs them into the folders Claude Code, Codex, Cursor and others read, and keeps them current at the start of every session.
- [Python Support for Runnable Code Servers](https://www.gatana.ai/blog/python-support-for-runnable-code): You can now write Runnable Code servers in Python 3.13. Put a schema dict and your functions in main.py, and Gatana deploys it and exposes the functions as MCP tools, the same way it does for Node.js.
- [How to Connect Claude to Your MCP Gateway](https://www.gatana.ai/blog/how-to-connect-claude-to-your-mcp-gateway): Point claude.ai, Claude Desktop, and Claude Code at one URL and every MCP server behind your Gatana gateway shows up in Claude. Step-by-step setup for OAuth and personal access tokens.
- [Tailscale VPN Support](https://www.gatana.ai/blog/tailscale-vpn-in-the-gateway): Executable Apps and Runnable Code deployed on Gatana can now join your Tailscale network. The server reaches the databases and internal APIs on your tailnet. Nothing on the tailnet can reach the server.
- [Persistent Storage in the Gateway](https://www.gatana.ai/blog/persistent-storage-state-that-survives-the-redeploy): In-gateway deployments of Executable Apps and Runnable FaaS can now keep a managed volume mounted at /data. Files survive restarts, redeploys, and idle stops, fully encrypted.
- [MCP's Largest Revision Yet Lands July 28](https://www.gatana.ai/blog/mcp-s-largest-revision-yet-lands-july-28): The 2026-07-28 spec makes MCP stateless, formalizes extensions, and hardens authorization. Here's what's changing, and what it means for everything behind your gateway.
- [Profiles: Programming the Identity Layer of Your Tool Gateway](https://www.gatana.ai/blog/profiles-programming-the-identity-layer-of-your-tool-gateway): The gateway between agents and your systems is where identity, credentials, and access control live. Profiles are how you make that boundary do exactly what your organization needs — per user, team, token, or IdP claim.
- [So why MCP? Honestly: it doesn't matter that much.](https://www.gatana.ai/blog/so-why-mcp-honestly-it-doesn-t-matter-that-much.): The value isn't the protocol — it's the governed boundary between agents and your systems. MCP is simply the standard everyone already speaks at that boundary.
- [Your REST API is Already an MCP Server](https://www.gatana.ai/blog/your-rest-api-is-already-an-mcp-server): Point Gatana at an OpenAPI spec and every endpoint becomes a typed MCP tool — no wrapper code to write or maintain.
- [Call Any MCP Tool Over HTTP](https://www.gatana.ai/blog/call-any-mcp-tool-over-http): Gatana's MCP HTTP API exposes every tool as an OpenAPI endpoint, so LLMs can orchestrate tools from code instead of pushing large payloads through the context window.

## Company

- Gatana c/o Accendo Consulting GmbH, Werdstrasse 122, 8003 Zürich, Switzerland
- Email: hello@gatana.ai
- GitHub: https://github.com/gatana-ai/ (TypeScript SDK and CLI, Python SDK, Helm chart)
- LinkedIn: https://www.linkedin.com/company/gatana-ai/
- Discord: https://discord.gg/mrEkuSUBvg

## Optional

- [Terms of service](https://www.gatana.ai/terms-of-service)
- [Privacy policy](https://www.gatana.ai/privacy-policy)
- [Data processing agreement](https://www.gatana.ai/dpa)