Hejdar MCP server

This is a small helper that connects your AI assistant to Hejdar, a service that checks actions against your organization's rules. Before your AI reads, writes, deletes, moves data or runs a command, it can ask Hejdar if that is allowed. It is handy for teams that want a safety net around what their AI is allowed to do.

What this MCP server does

You ask your AI to do something, like delete a customer record. The AI calls this helper and describes the action: what kind of action it is, which resource it touches, and any extra context. The helper sends that to Hejdar, which checks it against the rules your organization set up. Hejdar sends back a decision: allow, deny, or would deny. Your AI then either goes ahead or stops and tells you why.

Click to zoom

What you can do with it

• Check if an action is allowed before your AI does it
• Stop an action that your company's rules say no to
• See which rule was triggered and why
• See a risk level for the action
• Test rules in observe mode with would deny answers
• Pass extra details like department or reason for the check

Try asking your AI

• “Before you delete that customer record, check with Hejdar if it is allowed.”
• “Ask Hejdar if the hr-assistant agent may read the employee database.”
• “Check whether transferring funds from the finance account is permitted.”
• “Ask Hejdar if running this command on the server is allowed.”

What it gives back to you

You get back a short answer in the chat: allow, deny, or would deny. If it is denied, you also see the reason, like a rule that says deletions need manager approval. You may also see which rule was triggered and how risky the action is. Your AI then tells you what it found and stops if needed.

Before you start

Who it's for

Teams and admins who want a safety check on what their AI assistants are allowed to do, especially in companies with rules about data and actions.