Connect Claude or ChatGPT to Headrush, and your AI assistant can read and act on your modules, tasks, and learning targets — on your behalf, with your permissions, and only when you ask. This article walks through the connection and what to try first.
What is MCP?
MCP (Model Context Protocol) is an open standard that allows AI assistants like Claude to connect directly to tools and platforms — including Headrush. Once connected, your AI assistant can read and interact with Headrush data on your behalf, helping you save time on routine tasks.
What can you do with AI connected to Headrush?
Everything below respects your existing Headrush permissions — your AI assistant can only see and change what your account can already see and change.
With an AI assistant connected to Headrush via MCP, teachers and educators can:
Ask questions about student progress and learning targets
Generate module structures and task scaffolding
Search and browse learning modules without logging in separately
Get summaries of student activity and assessment data
Draft task descriptions, module goals, and learning objectives
Retroactively align targets to a student's past progress
Develop extension tasks to delve more deeply into a topic
Getting started
To connect an AI assistant to Headrush via MCP, you will need:
An active Headrush account with educator-level access
An AI assistant that supports MCP (such as Claude by Anthropic or ChatGPT)
Your Headrush MCP server URL. You can find this by clicking your Profile icon and looking for the "Connect Your AI Assistant" section:
Connecting Claude to Headrush
Open your Claude Desktop and click Customize > Connectors
Click the + icon and choose Add Custom Connector
Add Custom Connector details
Name: Headrush Learning MCP
URL:
https://yourschool.headrushapp.com/mcp— you can find your school's exact URL in your Headrush account settings.
Click Connect
Authenticate with your Headrush credentials when prompted.
Once connected, you can start asking Claude questions about your modules, tasks, and learning targets directly in the chat. A good test prompt: "List all my active modules"
Connecting ChatGPT to Headrush
ChatGPT supports MCP-backed connections through its Apps feature. To connect Headrush:
Open ChatGPT and go to Settings → Apps.
Browse the app directory or search for Headrush.
Click Connect and complete the authorization flow.
Once connected, you can mention Headrush in your prompts using @Headrush, or select it from the + menu in any conversation.
Note: App availability in ChatGPT depends on your plan. Some features may require a Plus, Pro, Business, or Enterprise subscription. If your organization is on an Enterprise or Edu plan, your workspace admin may need to enable the Headrush app first.
Prompt patterns that work well
The MCP responds best to prompts that are specific about what you want and grounded in your context. A few patterns we've seen work:
"Scaffold a [length] module on [topic] for [grade]. Use [methodology] as the design framework." — best for starting from blank.
"Look at the targets in [target set name] and propose alignments for [module name]." — best for retroactive alignment.
"Re-read the [task name] description and rewrite it for a student audience." — best for clean-up.
"Add a peer-critique checkpoint between [task group A] and [task group B]." — best for adding a beat to a unit you've already built.
Example prompts to try
"Show me a list of my active learning modules."
"Scaffold a new project-based module on environmental science for grades 9–10."
"What learning targets are aligned to my current module?"
"Help me write a task description for a research presentation task."
Current capabilities
We will be continually adding to the abilities of the Headrush MCP. Below are the actions that are currently available. Each action requires that you have the appropriate role on the module (owner, advisor, or admin).
Area | What the AI can do | Read / Write |
Modules | List your modules, view full module details, search across modules, create a blank module, scaffold a complete module with task groups and tasks, update module settings (title, description, dates, warehouse visibility, join settings) | Read & Write |
Learning Targets | Browse available target sets and their full target trees, align or remove learning targets on a module (cascades to tasks), align or remove targets on individual tasks | Read & Write |
Tasks | View full task details (description, due date, assessed targets, evidence count), create new tasks within a module, update task fields (title, description, dates, submission settings, assessed flag), move a task to a different task group | Read & Write |
Task Groups | List task groups (columns) within a module, create new task groups, update task group title, order, or learner visibility | Read & Write |
Site & Account | Retrieve school/site details (name, address, logo, base URL), retrieve your own user profile and role information | Read only |
Common issues
Claude says it can't see my modules.
Make sure you completed the auth flow with the same Headrush account you use day-to-day. If your school has multiple Headrush sites, you'll need to authenticate against the one your modules live on.
The AI scaffolded a module but the tasks are wrong.
Click Undo within ten minutes, then re-prompt with more specifics — grade level, unit length, learning targets, methodology. Vague prompts produce vague modules.
Can students use the MCP?
Not currently. MCP access is for educators with owner, advisor, or admin roles on a module.
Will my student data be used to train AI models?
No. We use models in inference mode only — your data is sent to the model provider only to answer your specific request and is not used for model training. See our Our approach to AI in Headrush article for the full picture.
Privacy and data access
When using AI with Headrush via MCP, the AI assistant only accesses data that your account has permission to view. No data is shared beyond what is necessary to respond to your request. Always follow your school's guidelines on AI tool usage and student data privacy.
Want to know more about how we think about AI?
Read Our approach to AI in Headrush — our principles, what AI in Headrush is good at today, and what it won't do.
Need help?
If you have trouble connecting or have questions about what your AI assistant can access, contact your Headrush administrator or reach out to Headrush support.