Cade turns any message, in a client's channel, a DM or your own dev channel, into a task or an incident in your dev channel, and follows each one through PRs, review, testing and client approval, keeping every thread and its notes for good.
Demo video coming soon
A message becomes a task, the team ships, the client approves. All in Slack.
Requests get lost between chats, trackers and inboxes. Cade keeps one thread per task, with no new tool to learn.
Clients, developers and QA all work in threads they already read. No new tool to log in to.
The team owns its tasks. The client approves only what they're asked to, and incidents stay urgent.
Suggestions are private and change nothing until someone clicks. Deletes ask twice.
Task cards and summaries are rendered from the same record, so they never disagree.
Requests arrive anywhere. Your team works in its dev channel. Cade brings each one in and keeps it moving.
On any message, in a client's channel or a DM: ⋯ → New task, pick the project, and the message is copied into the new task's thread.
Anywhere"@rex please start on this" offers to make Rex the owner and move it to In progress.
Dev channelPaste a PR and mention a reviewer. Cade links it and asks them to review.
Dev channelHand off to QA in plain words. Send finished PRs to the client early with @Cade release.
Client review asks whoever you pick to Approve or Not approve, in the task's thread.
Dev channelWrite "@mia can you review this" or paste a PR link in a task's thread. The proactive agent offers, privately, to assign or link it, optionally with AI, and changes nothing without a click.
/cade incident posts an urgent incident and tracks it in the dev channel immediately, marked as an incident. No triage step.
Send finished PRs to the client before the whole task is done. Whoever you pick approves or pushes back in the task's thread.
Stray messages are moved into the right thread with a link back, and marked 📤 so everyone can skip them.
/cade standup walks the team's open tasks one owner at a time. Update statuses and assign as you go.
Every reply in a task's thread is kept in Cade's database, with a page for the thread and one for notes, long after Slack's 90-day limit.
/cade … at the top of a channel, @Cade … in any thread, the ⋯ menu on task cards, message shortcuts and emoji. Commands always work, whatever else is switched on.
/cade taskTrack a task/cade incidentReport an incident@Cade status in reviewMove a task on@Cade pr add <link>Link a pull request@Cade moveMove a message into a thread/cade standupWalk the open tasksWrite “@mia can you review this?” or paste a PR link in a task's thread. Cade offers, privately, to assign or link it, optionally with AI, and changes nothing without a click.
Press Configure → AI agent on Cade's Home tab for an API key, and add Cade to Claude Code, Gemini CLI, Cursor or any client that takes a remote MCP server. The agent reads what you can read and changes tasks as you, after you approve.
claude mcp add --transport http cade https://YOUR-CADE-WORKER/mcp --header "x-apikey: YOUR_API_KEY"Video coming soon
Your own agent reading a task, proposing a change and posting it as you.
One command walks you through it: it asks for what it needs, runs each step, skips what's already done and tells you when something needs a click in your browser. No AI agent needed. Your tokens never leave your terminal.
Use the cade-deploy template on GitHub and clone your copy.
npm run setup starts the guided setup in your terminal.
Click through the logins and Slack settings when it asks. That's it.