MCP: Connect Edusign to AI assistants

The Edusign Model Context Protocol (MCP) server connects compatible AI assistants and development tools to Edusign.

The server is available at https://mcp.edusign.com and uses OAuth 2.1. When you connect, your MCP client opens a secure Edusign sign-in flow. You do not need to copy an API key or configure an Authorization header.

Watch the MCP setup video

What is MCP?

Model Context Protocol (MCP) is an open standard that lets AI applications securely access external tools and data. With the Edusign MCP server, a compatible assistant can:

Work with live Edusign data and tools
Respect the permissions of the authenticated Edusign user
Help with courses, attendance, groups, learners, documents, and other supported Edusign workflows
Use OAuth instead of exposing an API key in client configuration

Available tools

Compatible clients discover these tools automatically through MCP. The names below are the stable tool identifiers that agents can use when planning a request.

School and learning workflows

These high-level tools are read-only and optimized for common Edusign tasks:

Tool	What it provides	Useful discovery hint
`get-school-context`	Connected school identity, timezone, language, and absence types	Call first so dates and absence filters use the school's configuration
`search-students`	Students matching a name or email fragment	Use before requesting one student's detailed overview
`get-student-overview`	Student profile, courses, justified absences, and assessment submissions	Requires a student ID returned by `search-students`
`list-courses`	Courses filtered by period, professor, student, group, training, or name	Keep filters focused when querying large periods
`get-course-attendance`	Attendance sheet with presence, signatures, delays, justified absences, and totals	Requires a course ID returned by `list-courses`
`list-absences`	Justified absences filtered by date, student, course, or absence type	Use `get-school-context` to discover absence type IDs
`get-training-overview`	Training details, enrolled people, and related courses	Use when an agent needs the complete context of one program
`list-professors`	Professors, optionally filtered by name or email	Useful for resolving a professor before filtering courses
`list-groups`	Groups and, optionally, their students	Useful for resolving a class before filtering courses

API discovery and execution

These tools let an agent discover the documented Edusign APIs before making a request:

Tool	What it provides
`get-api-guide`	Edusign API conventions for versions, pagination, dates, identifiers, errors, rate limits, and access
`list-specs`	API specifications available through the MCP connection
`list-endpoints`	Operations in one specification, optionally filtered by tag
`search-endpoints`	Full-text search across endpoint paths, summaries, tags, and descriptions
`get-endpoint`	Parameters, request and response schemas, errors, and examples for one operation
`execute-request`	Executes a documented operation with the connected Edusign account; authentication is handled by the server

execute-request is limited by the OAuth scopes and Edusign permissions of the connected user. A read-only connection refuses write operations. When write access is granted, agents should inspect the endpoint first and ask for explicit confirmation before destructive actions.

Prompts and resources

The server also exposes two reusable prompt templates:

attendance-audit — builds an attendance report for a training and period
student-review — builds a complete learner review for a meeting or follow-up

Agents can also discover compact API specification resources through MCP. These resources help them inspect the current, legacy, and partner API surfaces without guessing endpoints.

Recommended agent workflow

Call get-school-context at the beginning of a business-data session.
Prefer the high-level school and learning tools for common questions.
For another documented API operation, call get-api-guide, then search-endpoints or list-endpoints.
Inspect the selected operation with get-endpoint before calling execute-request.
Respect the connected user's permissions and confirm any destructive action.

Server details

Setting	Value
Server URL	`https://mcp.edusign.com`
Transport	Streamable HTTP
Authentication	OAuth 2.1

Connect your MCP client

For ChatGPT, Claude, Cursor, or another compatible client:

Open the client's MCP, connector, or integration settings.
Add a custom MCP server named Edusign.
Enter https://mcp.edusign.com as the server URL.
Save the configuration.
Complete the Edusign OAuth sign-in flow in your browser.

Each user signs in with their own Edusign account. The assistant can only access data and actions allowed by that user's permissions.

Using ChatGPT on the web? Add Edusign through ChatGPT's Plugins or Connectors interface. ChatGPT web does not read MCP configurations installed from the terminal.

Add the server to ~/.cursor/mcp.json:

{
  "mcpServers": {
    "edusign": {
      "url": "https://mcp.edusign.com"
    }
  }
}

Restart Cursor if needed, then complete the OAuth sign-in when prompted.

Test the connection

After authentication, start a new conversation and try a simple request such as:

"List my three most recent courses in Edusign."
"Show the justified absences for this month."
"Which attendance sheets are still waiting for signatures?"

If the assistant returns data that matches your Edusign account and permissions, the connection is working.

Security recommendations

Sign in only through the secure Edusign OAuth flow.
Never paste Edusign credentials, API keys, or access tokens into a prompt.
Use a trusted MCP-compatible client.
Disconnect and authenticate again if you have any doubt about the connection.

For a guided setup, see the Edusign Help Center article.