Quick answer: Open your AI agent, give it your Eduqat API key, and tell it in one sentence to connect to the Eduqat API. It reads the documentation itself and starts working. After that you ask for what you want in plain language.

The Eduqat API exposes your school to anything that can send an HTTP request. Pairing it with an AI agent means you never write those requests yourself.

Once connected, you can ask things like how many self-paced courses do I have, and which are still draft, or say create a self-paced course called Intro to Design Thinking with three sessions. The agent works out which calls to make, shows them to you, and runs them once you approve.

Typical uses are building courses, sessions, and materials, enrolling learners and checking certificates, and pulling lists out of your school for a report.

This is different from the AI School Assistant. The assistant lives inside your Eduqat Dashboard and needs nothing set up at all. The API is for work outside the dashboard, and for anything you want to script, export, or repeat.

That is the whole list. You do not need to know how to code, and you do not need to touch a terminal to get started.

Ask your Eduqat account manager for an API key for your school. Keys are issued per school, and there is no way to generate one yourself from the dashboard.

Two things to know before it arrives:

The key identifies your school on its own. You never send a school id, a user id, or a role alongside it. The server works all of that out from the key.

The key is single tier. There is no separate read only key and no separate write key, so one key grants every operation the API exposes. Treat it like a password.

Open your agent and hand it the key. You can attach the key as a file or paste it into the message, whichever your agent supports.

Then send this.

That is the whole setup. The agent reads Eduqat's guide for AI agents, finds the machine readable specification linked from it, makes a test call to confirm the key works, and tells you what it found.

Pointing the agent at https://developers.eduqat.com/docs works too. It will find its way to the same guide from there.

Once the agent reports a successful connection, stop thinking about endpoints and just describe the job.

Reading and writing both work this way. The agent decides which calls to make, asks permission before each one, and tells you what it did.

From there you can keep going in the same conversation. Asking it to turn an answer into a spreadsheet, or to clean up duplicates it found, is a normal next request.

This is the step most people skip, and it is where wrong answers come from.

Watch the agent's tool calls. Before any API call, you should see it fetch the guide, something like this.

If no fetch happens and the agent goes straight to building requests, it is guessing from its training data rather than reading the current specification. Stop it and say so.

Everything above is enough for a one-off job. Two extra steps make repeat work safer and faster. Neither is required.

Keep the key out of your chat history. Pasting the key into a message means it stays in that conversation. Storing it as an environment variable instead lets the agent use it without the value ever appearing in the chat.

On macOS, add this to ~/.zshrc. On Linux, add it to ~/.bashrc.

Then reload the shell and confirm it is set.

On Windows, set it once from Command Prompt or PowerShell, then open a new terminal so the change takes effect.

Now tell the agent to use $EDUQAT_API_KEY instead of giving it the key directly.

Make every new session start ready. Create a folder for Eduqat work and put a project instruction file in it. Every future session you open in that folder is configured before you type anything.

In Claude Code the file is CLAUDE.md. Ask your agent to create it with these contents.

Codex, Cursor, and other agents each have their own context or instructions file, so check your agent's documentation for where it lives. The contents are the same three facts every time: the URL https://developers.eduqat.com/AGENTS.md, the variable name EDUQAT_API_KEY, and the header name x-api-key.

Check a key yourself. If you want to confirm a key before handing it to anything, run this in a terminal and replace YOUR_API_KEY.

JSON with an items array means the key works. 401 means the key is wrong, and 403 means it is valid but lacks permission for that route.

The key grants every operation on your school, so the usual rules apply with no exceptions.

Never commit it. If your agent writes the key into a .env file, keep that file out of git.

Never paste it into a shared chat, a ticket, or a document other people can read. A private session with your own agent is fine, but anything other people can open is not.

If you think it leaked, tell your account manager straight away and ask for a new one.

For anything about the key itself, including a 403 that will not go away, go to your account manager. For a documented path that returns 404, or anything that looks like a bug, reach out through the live chat button in the bottom right corner of your screen and we will look into it with you. The full developer documentation lives at https://developers.eduqat.com/docs.