The MCP gateway that makes Your organization,
AI Ready.
Centralize and share skills, prompts, and tools.
Coding-agents, chat-bots, A2A, assistants; all can use Gatana.
Works with the agents your team already uses
What is Gatana?
One gateway between your agents and your tools
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.
Everything-in-one
Tools. Skills. Prompts.
One place.
Built to support complex agentic scenarios. Use Gatana for your entire organization, separate between teams and departments.
Start Free TrialConnect any MCP, be it HTTP or local executable e.g. uvx or npx. Persistent storage supported.
Convert OpenAPI or Swagger to tools. Either upload your own OpenAPI spec to Gatana or point to the JSON URL.
Deploy FaaS serverless code on Gatana. Connect to any tool or service, even if it's missing an OpenAPI spec or MCP. Persistent storage supported.
Attach a managed storage volume to enable durable storage for databases and long-term memories.
Built-in support for Tailscale. Seamlessly connect private services inside your tailnet
Create and share skills with persons or teams. Gatana installs a hook to keep them in-sync.
Agents publish reports, dashboards and small tools at a stable link. Share them with people, teams, or anyone with the link.
Access tools over HTTP instead of MCP. Gatana provides well-specified OpenAPI specs for all tools.
Extremely Flexible
Control identity and access
Built to support diverse and complex agent landscapes.
Precisely control access for your entire organization.
Powerful credential and access controls
Profiles help you solve complex scenarios.
Magic Auth Link
Single-use links that authenticates
Auth for all services
Supports OAuth 2.1, API keys, headers, URL keys.
Customize anything that can connect
Accounts, teams, API keys, OAuth clients, JWT claims. All can be customized.
Engineering Focus
A platform for all developers AI needs
Use the shared tools and skills to quickly onboard your team-members. Keep sensitive credentials secure and away from local hard drives.
- Profiles for complex access control scenarios
- Integrate Gatana in your app with the SDK
- Well-typed APIs, JSON responses, OpenAPI spec
- On-Premise Deployment
from langchain_mcp_adapters.client import MultiServerMCPClient
client = MultiServerMCPClient({"gatana": {
"transport": "streamable_http",
"url": "https://ORG_ID.gatana.ai/mcp",
"headers": {"Authorization": f"Bearer {PAT}"},
}})
tools = await client.get_tools()claude mcp add gatana --transport http https://ORG_ID.gatana.ai/mcp
# Sign in once in the browser. Every tool you are allowed is now in Claude.codex mcp add gatana --url https://ORG_ID.gatana.ai/mcp
# Sign in once in the browser. Every tool you are allowed is now in Codex.hermes mcp add gatana --url https://ORG_ID.gatana.ai --auth oauth
# Sign in once in the browser. Every tool you are allowed is now in Hermes.// ~/.openclaw/openclaw.json
{ mcp: { servers: { gatana: {
command: "npx",
args: ["-y", "mcp-remote", "https://ORG_ID.gatana.ai/mcp"]
} } } }Federated identity
SAML, OIDC, SCIM? No sweat.
Sync people and teams from Okta, Entra ID, or any OIDC and SAML provider.
Built for the security team
Prove what every agent did.
Every tool call is recorded: who asked, which agent acted, on whose behalf, and what it touched. Stream it to your SIEM, gate it with a firewall, and keep the secrets in your own vault.
Audit logs
Every call, every credential use, every admin change, with the person and the agent behind it.
HMAC-signed NDJSON over HTTPS to Splunk, Falcon LogScale, Elastic, Sentinel, or any collector.
Allow or block tools per user, team, or profile, so an agent only sees what it is meant to use.
Reference secrets in AWS, GCP, HashiCorp Vault, Infisical, or Azure Key Vault instead of copying them.
Ephemeral Ubuntu containers for autonomous agents, isolated in Kata VMs, with the Gatana CLI installed.
Gatana Cloud runs in the EU. Or self-host on-premise for full control over residency and compliance.
Why Gatana?
Expand what your team can do
Giving AI real access used to mean API keys pasted into chat windows, one-off integrations nobody maintains, and no answer to “what did the AI do?”. Rollouts stalled at the security review.
With Gatana, every person and every agent gets governed access from day one: the tools they should have, under their own identity, with a record of everything.

Read
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.
99.7%
of public OpenAPI and Swagger specifications converted into MCP tools, every operation a tool.
29 ms
or 3%, of a typical tool call is gateway overhead.
44×
fewer prompt tokens with code mode on the largest task: 30,458 against 1.33 million for direct tool calls. 4 times fewer on the other large task.
Measured 5 October 2026
How we measured, with the raw dataPricing
No credit card for the free plan. Cancel the Pro trial at any time.
Pro
Most popular$125/ month
Gatana Cloud, EU data residency.
Start 14-day free trial- Unlimited users
- Unlimited servers
- Federated identity
- 20 GiB storage
Enterprise
Custom
On-premise, on your terms.
Schedule a call- Customizations
- Dedicated support
- Engineering consulting
- Enterprise integrations
- On-premise
FAQ
Questions, answered.
The short version of what people ask before they connect their first agent. The docs have the long one.
What is Gatana?
- Gatana is an MCP gateway that offers turn-key support for skills, SSO, and more. Your organization can connect their services, REST APIs and serverless functions, and every approved AI agent can use them under the user’s own identity, with credentials, access control and audit logs handled in one place.
What is an MCP gateway?
- An MCP gateway is a server that sits between AI agents and the services they use. It offers governance, and mananages authentication, access control, credential storage and audit logging.
Which agents and systems work with Gatana?
- Any agent that supports MCP: Claude (claude.ai, Claude Desktop and Claude Code), ChatGPT, Codex, Cursor, Windsurf, VS Code and GitHub Copilot, Gemini, Zed, JetBrains IDEs, OpenCode, Cline, n8n, LangChain and other agent frameworks.
Can I connect a REST API that has no MCP server?
- Yes. Point Gatana at an OpenAPI or Swagger specification and every endpoint becomes a typed MCP tool, with no wrapper code to write or maintain. For a service without a specification, write a small Runnable Code function in Node.js or Python and Gatana deploys it as a tool.
Does Gatana support SSO, SAML and SCIM?
- Yes. Gatana federates identity with Okta, Microsoft Entra ID and any SAML or OIDC provider, provisions people and teams through SCIM, and supports just-in-time onboarding from federated tokens. Federated identity is included in the Pro and Enterprise plans.
Can I self-host Gatana?
- Yes. Enterprise customers run Gatana on their own Kubernetes cluster with the official Helm chart, which gives full control over data residency and compliance. Gatana Cloud is the hosted option; it runs in the EU and is used by the Free and Pro plans.
How much does Gatana cost?
- Free costs $0 and covers one user and three enabled servers, remote MCP or OpenAPI. Pro costs $125 per month per organization, with unlimited users and servers, local and Runnable Code servers, federated identity, SIEM streaming and 20 GiB of storage, after a 14-day trial with no credit card. Enterprise is custom and includes on-premise deployment.
What does Gatana record for audit?
- Every tool call, credential use and admin change, together with the person who asked, the agent that acted on their behalf and the server it touched. The log can be streamed as HMAC-signed NDJSON over HTTPS to Splunk, Falcon LogScale, Elastic, Microsoft Sentinel or any other collector.
From the blog
Latest from Gatana
Code Mode vs. Direct Tool Calls: The Same Three Tasks, Five Times Each
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? We Measured Ours: 52 Milliseconds, Then 29.
We timed 10,800 tool calls, direct to a server and through Gatana, from inside the same cluster. In the morning the gateway added 52 ms per call; after a fix the same afternoon, 29 ms. The setup, the raw numbers, where the time goes and what we make of it.

We Converted 300 Public OpenAPI Specs to MCP Tools. Here Is What Broke.
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.
Your team, AI Ready.
Free for individuals. A 14-day Pro trial for teams, no credit card. Or talk to us about running Gatana on your own infrastructure.
