Model Context Protocol

Connect Claude, Cursor or your own agent to Kantex Projects via MCP

Kantex Projects has its own MCP server. Claude, Cursor or your own agent can search tickets, create bugs, post comments and link branches. For each token you choose whether the agent may read, write or delete.

Example conversation with an AI assistant connected to Kantex via MCP

Vraag: Which critical bugs are open for Supermarkt Asperen? Antwoord: There are 2 critical bugs open in ASUP: ASUP-14 (payment fails with iDEAL, branch fix/ASUP-14-ideal-callback, pull request open) and ASUP-17 (prices not visible on mobile, no branch yet).

via projecten.kantex.nl/api/mcp · list_bugs
Connected in 3 minutes

Choose your AI client

In Kantex Projects, create a token under Integrations › MCP. Choose which permissions the agent gets: read, write or delete.

  1. In Kantex Projects, create a token under Integrations › MCP.
  2. Run the command below and replace kp_... with your token.
  3. Ask Claude Code about your tickets, for example: what is open in ASUP?
Terminal
claude mcp add --transport http kantex-projects https://projecten.kantex.nl/api/mcp \
  --header "Authorization: Bearer kp_..."
Tools

What your AI client can do

  • list_spacesReadSpaces (projects) with their client
  • list_issuesReadFind tickets by space, status, type or text
  • get_issueReadA single ticket with subtasks, branches and bug details
  • list_bugsReadOpen bugs, sorted by severity
  • list_labelsReadThe label structure (group:value)
  • create_issueWriteCreate a ticket in a space
  • create_bugWriteReport a bug with severity, environment and steps
  • update_issueWriteChange status, title or labels
  • add_commentWritePost a comment on a ticket
  • link_branchWriteLink a branch to a ticket
  • run_code_scanWriteScan a diff or code snippet
  • delete_issueWriteDelete a ticket (only with the delete permission)

Permissions per token

Read, write or delete, and individual tools can be switched off. Revoking takes effect immediately.

Same permissions as the team member

A token never sees more than the team member who created it. Clients cannot create tokens.

Hosted in the EU

The MCP server runs inside Kantex Projects itself, on servers in the EU.

Frequently asked questions

Questions about AI and MCP

Which AI clients can I connect?

Any MCP client that supports the streamable HTTP transport with a bearer token, such as Claude Code, Claude Desktop (via mcp-remote) and Cursor. Your own agent can call the server too.

What can an agent do with my token?

You choose that per token: read, write or delete. Without the delete permission, the server refuses delete_issue. You can also switch off individual tools per token and revoke a token immediately.

Does the agent see everything?

A token works with the permissions of the team member who created it, and never with more. Clients in the client portal cannot create tokens.

What is KantexAI?

KantexAI is the assistant built into Kantex Projects. It answers questions about your tickets, turns a rough report into a complete bug report and suggests severity and labels during triage. KantexAI is included in Business and can be added separately for €19 per month.

What does the MCP server cost?

The MCP server is included in every plan, including Free.

Try it with your own data

MCP is part of the AI agent & MCP connection module, and you can try it free for 14 days.

  • No credit card
  • Cancel monthly
  • Support from the Netherlands