Claude Code
One command adds the server; the plugin adds the Always lines and the session hooks. Several sessions at once, each on its own claim and worktree.
Claude Code setup →Integrations
Connect the tools you already pay for.
Claude connects over MCP. Your boards are read by your agents, as you. Your design files file their own tasks. Each integration has a real setup page, not a swapped noun.
Claude
Three ways in. One board.
Every Claude acts for a named person and shows as "Claude for Noor". Agents never need a seat.
claude.ai
Add a custom connector with the MCP URL, approve it once, and work the board from the browser: plan, ask, record decisions.
claude.ai setup →Claude Desktop
The same connector URL. Desktop sessions get the briefing too and can hold a claim like any other chat.
Claude Desktop setup →External boards
Keep the client's board. Work in Projectify.
An agent reads the board as you, cards become tasks that read like the card, and a conflict shows on the task and on the board, where you already look.
Trello
Add the board link and a note; an agent reads the cards as you, no token to paste. Lists become statuses; cards read like the card.
Trello setup →Monday
Groups and columns map to statuses; items, subitems, updates with replies and their files sync across.
Monday setup →Azure DevOps
Work items become tasks, parents become subtasks, states become statuses. An agent reads them as you.
Azure DevOps setup →Design and code
Files that file their own tasks.
Figma
Link a file and say what to look for. An agent reads it and files what it finds as tasks, each finding once.
Figma setup →GitHub
The repo is matched to its project from the remote. Claims carry the worktree and branch; the outcome carries the PR.
GitHub setup →Routines
A routine wakes Claude on an event: an answer you gave, or a task done, so the next one is picked up. Your Claude Code routine runs it in the cloud, capped and attributed.
How routines work →