Docs / Integrations / MCP server

Manage Kevta from your AI assistant

Connect Kevta to Claude Code, Claude Desktop, or any MCP-compatible client and run your tasks by just asking — create issues, update status, leave comments, close things out. No context switch, no copy-paste.

Overview#

MCP (the Model Context Protocol) is the open standard AI tools use to talk to the apps you already run. Kevta speaks it, so your assistant can act on your boards directly — always within the permissions you already have.

“Create a bug ticket for the slow login issue on the Web board.”
“What's still open and assigned to me this week?”
“Move KEV-42 to In Review and comment that the fix is deployed.”

How it works#

Three steps, once. After that, you just talk to your assistant.

  1. Generate a token in Kevta. A token is how your AI client proves it's you. Create one in your settings and copy it.
  2. Add Kevta to your AI client. Paste the token and Kevta's address into Claude Code or Claude Desktop once.
  3. Ask in plain language. Your assistant finds the right workspace, board, and task and makes the change — it appears in Kevta instantly.

Get a token#

Tokens come in two kinds, depending on how much access you want to hand over. Both expire, both can be revoked anytime, and both act as you.

All access

User-wide token

Reaches every workspace and board you can. Best for your own daily use.

  1. Open Profile → API tokens (MCP).
  2. Name it and pick 7 / 30 / 90 days.
  3. Click Create token and copy it.
Board only

Board-scoped token

Limited to a single board. Best for a bot or a shared, narrow task.

  1. Open that board's Settings → API tokens.
  2. Name it and choose an expiry.
  3. Click Create token and copy it.
!
Copy it right away. The token is shown once, at creation, and never again. If you lose it, revoke it and create a new one.

Connect your client#

Point your client at https://api-kevta.matrixmindsit.com/mcp with your token as a bearer header.

Claude Code
# add Kevta as a remote MCP server
claude mcp add --transport http kevta https://api-kevta.matrixmindsit.com/mcp \
  --header "Authorization: Bearer YOUR_TOKEN"
Claude Desktop · claude_desktop_config.json
{
  "mcpServers": {
    "kevta": {
      "type": "http",
      "url": "https://api-kevta.matrixmindsit.com/mcp",
      "headers": { "Authorization": "Bearer YOUR_TOKEN" }
    }
  }
}
i
There's no sign-in pop-up to click through — pasting the token is the whole setup. That's expected, not a missing step.

What you can ask for#

Nine tools your assistant picks from automatically. You never need to know workspace or task IDs — it looks them up.

Read
list_workspacesSee the workspaces you belong to.
list_projectsSee the boards inside a workspace.
describe_boardSee a board's own fields — due date, type, any custom column — with their valid values.
list_tasksList a board's tasks, filtered by status or assignee.
get_taskRead one task in full, including its comments.
Write
create_taskAdd a task with title, description, priority, assignee, due date — or any other field the board defines.
update_taskChange title, description, status, priority, assignee, due date, or any other field the board defines.
add_commentLeave a comment on a task.
close_taskMove a task to its board's done status.

Safe by design#

A token never grants more than you already have, and access is checked fresh on every request.

  • Acts as you, live. Every action is checked against your current Kevta permissions. Lose access to a board and your token loses it too — instantly, no reissue.
  • Scoped when you want. A board-scoped token can only ever touch that one board, even if handed another board's ID.
  • Expiring & revocable. Every token has a fixed 7/30/90-day life and can be revoked from settings anytime.
  • Stored hashed, never logged. Tokens are kept only as a one-way hash, and task content never appears in server logs.
✓
Manage or revoke your tokens anytime in Profile → API tokens (MCP) or a board's Settings. See API tokens for the full reference.
Kevta is in beta