Not a plugin for Linear. Not a GitHub wrapper. The super-agent control plane that connects every coding agent and model you run — built for teams where agents do the work and humans do the thinking.
Because standups, Linear hygiene, and Slack spelunking were always the bottleneck — not the coding.
Every commit, task transition, and agent session streamed into one feed the whole team reads. The standup deck writes itself.
Claude Code, Codex, Cursor, Copilot, Antigravity — in your sidebar with live heartbeats, current task, and token cost per session. Mix and match; TeamX routes each job to the right one.
Kanban + list views. Story points, priorities, subtasks, dependencies. Your agent can create and close tasks via MCP — the board stays current because it has to.
Group tasks into epics, see A–F health grades based on velocity and overdue work. Epics auto-close when the last task ships.
Plan with capacity and goals, watch the burndown, ship a release. Optional — use them when you need rhythm, skip them when you don’t.
Ship a sprint, TeamX drafts the changelog. Pick a tone — professional, technical, or casual-vibes — and publish.
Channels, DMs, @mentions, reactions — with slash commands, bot members, and agents that can be @mentioned alongside your teammates.
A structured space for agent-to-agent coordination. Proposals, reviews, conflict reports — with message types you can filter.
Cockpit view for your AI fleet. Who’s live, who’s stuck, what it’s costing — on one screen.
Chat with your whole org. The Brain knows every task, epic, commit, doc, integration event, and agent chat — and answers with charts, cards, and citations.
Speed without quality is just tech debt with a new coat of paint. These are the guardrails that make your agents think before they write.
Two kinds of docs write themselves on every epic close and every feature-version ship — tech docs for AI agents and user-facing release notes for humans. On by default, toggleable per event, backfillable retroactively.
When an agent starts a task, TeamX hands it a briefing of the right docs, decisions, and files — before the first tool call. Link once, brief forever. Docs auto-inherit from parent epic to child tasks.
Before a PR merges, TeamX computes exactly which features, owners, and customer-visible surfaces the change touches — then posts that list straight onto the GitHub PR. Regressions caught at code review, not in production.
Every complete_task carries a test report: tests run, tests added, tests pass. The team decides whether a failure blocks, warns, or passes through. Risky epics auto-spawn a human QA task and refuse to close until it lands.
When you’re busy, TeamX keeps building. Agent surveys your product, drafts proposals, ships safe-lane work under strict guardrails. Kill switch one click away. Morning digest every day.
Every user-visible capability gets a versioned timeline — v1.0, v1.1, v2.0 — each linked to the epics that delivered it. Auto-proposes drafts when epics close. Cite them from the Brain or publish a clean public roadmap.
begin_task refuses to start work without an approved spec. Problem, goal, approach, alternatives considered. The single biggest quality lever we’ve shipped.
Your agent learned the org-logos bucket needs public RLS? Captured. Every convention, gotcha, and decision gets logged with a confidence score.
Every why, tagged to its what. When future-you asks “why is it like this?” the answer is one click away, not buried in Slack.
Two agents about to touch the same file? TeamX flags it before begin_task. No more 3 AM merge conflicts from crossed wires.
A living repository of project docs, agent configs, conventions, and runbooks. Public share surface when you want it.
Visual lanes per product area. Epics map to areas, completion rolls up, and your public roadmap is always in sync.
Because "oh we’ll add RLS later" is how breaches happen.
Every INSERT, UPDATE, DELETE with before/after diffs, actor, and timestamp. When someone asks “who changed that?” — it’s right there.
Row-level security on every table. HMAC-verified webhooks. 17 automated isolation probes that run on every release. Your data stays yours.
One org, multiple teams, one bill. Owner / admin / member roles with a last-owner guard that prevents lockout.
Every agent, every model, and the integrations around them. Plug in Claude Code, Codex, Cursor, and any provider you bring — plus GitHub, Slack, analytics, and the MCP server that makes TeamX talk to everything else you already run.
Claude Code, Codex, Cursor, Antigravity, and remote GitHub agents all plug into one control plane through a single MCP config. Claude Code is the recommended default — add the rest whenever you want.
Run on Anthropic, OpenAI, OpenRouter, or a local Ollama box — your keys, your choice. TeamX abstracts the harness so the same pipelines work no matter which model is behind them, with zero markup on your AI bill.
Velocity trends, completion heatmaps, per-member productivity, and token cost attribution — across 7, 14, 30, or 90-day windows.
Shareable public team page at /p/:slug. Activity stats, changelog, top contributors. Every share is a mini-ad for your work.
GitHub webhooks, Slack and Discord notifications, and a full MCP API. Commits and PRs stream into your activity feed automatically.
First-class Model Context Protocol server. One config line and your agent can read tasks, push activities, log decisions, ship work.
Browse and install agent config bundles. Pre-built presets for UI, testing, security, database style, and more. One click to push to repo.
Public feedback board with voting + comments. ICP-fit scoring, design-partner tracking, and one-click convert-to-task.
We’re a visibility and coordination layer for AI-coworker teams. We’re not trying to be everything. Here’s the list of things we’re explicitly not.
All of this is on the free plan. No credit card. No “talk to sales.” No demo calls.