Turn Your Stack9 App Into Tools for AI Agents
You already modeled your data as entities and your business logic as automations. Stack9 Instance MCP exposes the query-library queries and automations you choose as secure MCP tools — so any MCP-capable AI agent can read your data and trigger your workflows, under the same permissions your REST API already enforces.
Your App's Logic, Ready for AI Agents
Pick queries and automations, get tools
Choose entity-backed query-library queries and webhook-triggered automations to expose. Stack9 auto-derives input schemas from your generated models — no bespoke integration code to write.
Author MCP servers visually
Build and manage MCP servers from App Builder → MCPs in the Console: name a server, add tool cards, and test them live against a real MCP session before you ship the URL to anyone.
Enterprise-grade, fail-closed authorization
Every tool call is re-authorized inside the request: query tools require entity read privilege, automation tools require automation permission. Anything not explicitly allowed is denied — never left open by omission.
Works with the MCP clients your team already uses
The endpoint speaks MCP Streamable HTTP, so MCP-capable clients — including Claude — can connect, list tools, and call them, authenticating with a Stack9 session cookie or API key.
Read data and take action, from one surface
Query tools give agents safe, read-only access to your entities. Automation tools let an agent trigger real side effects — create a ticket, kick off a workflow — through the exact automation you already built.
No duplicate data layer for AI
Instance MCP is built on the entities, permissions, and query library you already have. There is nothing new to model and nothing separate to keep in sync.
How Instance MCP Works
Pick your queries and automations
In Console → App Builder → MCPs, create an MCP server and add tools — each backed by an entity-backed query or a webhook-triggered automation.
Get a secure MCP URL
Stack9 generates one endpoint per server: {coreBaseUrl}/api/mcp/{mcpKey}. Copy it from the "Get MCP URL" action.
Connect your AI agent
Point any MCP-capable client at that URL over Streamable HTTP, authenticating with a Stack9 session cookie or API key — the same credentials your platform already issues.
Agents call tools, Stack9 enforces the rules
Every call is checked against real entity read privileges and automation permissions before it runs. No permission, no result — fail closed, every time.
See It in the Console
Author, test, and ship MCP servers without leaving Stack9 Console.
List and manage every MCP server your instance exposes.
Add tools from the Queries or Webhook Automations group, with a live "Generated model found" badge.
Connect a real MCP session and run a tool call before you share the URL.
Copy the endpoint URL and hand it to your AI agent's MCP client configuration.
Make Your Stack9 App Agent-Ready
Every entity, query, and automation you've already built is a candidate MCP tool. Authenticate with a Stack9 session or API key — there's no separate credential to provision.