Developers
Integrate Elevated Context
Elevated Context is an authorization and information-flow governance plane for AI. It decides what context an AI is allowed to receive for the task in front of it — never more than the person asking is permitted to see. These pages show how to connect to it.
Choose your integration path
Elevated Context MCP
A local Model Context Protocol tool that asks Elevated Context for governed context on behalf of a signed-in person. Best if your AI client speaks MCP.
On requestGovernance-safe client
A thin TypeScript client over the governed request protocol. The server stays the sole authority; the client carries none. Best if you are building your own tool.
Limited (TypeScript)Governed request API
The underlying request contract with a canonical OpenAPI artifact. Best if you want to understand the wire protocol and governance outcomes directly.
Limited (private)Not sure which one?
- Your AI assistant supports MCP (for example, an MCP-capable desktop or IDE client): start with Elevated Context MCP. It is the most complete supported path today.
- You are writing your own TypeScript tool or service: use the governance-safe client, which wraps the same protocol.
- You want the raw contract and outcome model: read the governed request API reference.
How a request is governed
Every integration path funnels into the same governed decision. Elevated Context establishes who is asking, confirms what they are permitted to see, narrows to what the purpose actually needs, and releases only that — or holds back and fails closed if it cannot confirm the permission a release would need. A client never selects its own authority; the server decides.