Klarify AI Overview
Klarify AI exposes your organization’s processes, task instructions, and people data to AI clients through the Model Context Protocol (MCP). Once connected, an AI assistant can search documents, summarize processes, and—for users with the right role—create or update content on your behalf.
Who this is for
Klarify AI documentation is written for the technical person setting up the integration: an admin, IT staff member, or automation builder configuring an AI client to talk to Klarify. End users of the AI client itself do not need to read these pages.
Supported AI clients
Klarify AI works with any MCP-compatible client. The setup guides cover:
Other MCP clients can connect using the same endpoint and authentication flow.
What an AI client can do
Once connected, an AI client can take actions on your behalf — searching for a process model, listing employees in a department, creating a new position, and more. See the tool reference for the full list.
What an individual user can do through the AI client depends on their Klarify role. Read-only access is available to all members, while creating, updating, and deleting content requires an administrator role. See Permissions and roles for the full breakdown.
To discover what’s available in a given session, ask the assistant directly — for example, “What can you do?” It returns a complete, role-filtered summary of the actions it can take for you.
What the AI won’t do
The assistant is scoped to your organization’s data — its people, departments, positions, teams, regions, locations, processes, task instructions, and documents. It is not a general-purpose assistant, so it politely declines requests with no connection to your organization, such as general world facts, current events, weather, math, coding help, or creative writing.
It also won’t guess at questions it has no grounded data for. Klarify does not yet hold a defined source for an organization’s mission, value proposition, value chain, strategy, or competitive advantage, so the assistant declines these rather than inventing an answer. It can still summarize how your organization is structured and what it does — but only from real data such as your departments, positions, and documented processes, framed explicitly as based on what’s documented in Klarify.
Prerequisites
- An active Klarify organization
- A Klarify account in that organization
- An MCP-compatible AI client (Claude Desktop, Microsoft Copilot Studio, or similar)
Klarify AI is available to every Klarify organization. There is no separate provisioning step or plan upgrade required.
How authentication works
When an AI client connects for the first time, Klarify opens a browser window so you can sign in with your existing Klarify credentials. After you sign in, the AI client receives an access token that it stores locally and uses for subsequent requests. The token is tied to your Klarify user and selected organization, so the AI client inherits exactly the permissions you have in the app.
Connection details
| Detail | Value |
|---|---|
| Endpoint | https://mcp.klarify.biz/mcp |
| Transport | Streamable HTTP |
| Authentication | Browser-based authentication |