Twindem Suite · Team — private preview

The terminal your AI agents work in.

Claude Code and Codex run unmediated, exactly as in your favorite terminal. Team adds what a bare terminal can't: every tab is a board task, agents start briefed, a second agent reviews with a typed verdict — and nothing ships while findings are open.

Native macOS, signed & notarized, auto-updating; Windows next
One sign-in — projects, board access and agent tools flow from your Twindem account
Works with the agents you already use: Claude Code and OpenAI Codex

Request early access How it connects

DB-0142 auth-fix · In progress · CI ✓ · A1 idle 2m

> claude "/twd-context DB-0142"

⏺ Reading the board: brief, history, approved designs…

⏺ Implementing per the approved architecture (DB-0101)…

⚑ Review verdict: request_changes · 2 findings → blocked from Done

One platform, two native clients

Team builds. Pulse communicates. The platform remembers.

Everything meets on twindem.ai — Admin (organization, projects, seats), the Board (the single source of truth for work) and Docs (architecture, decisions, approved designs). Twindem Team is where humans and agents build; Twindem Pulse is where the project communicates. Same account, same projects, same truth.

Build

Twindem Team

Terminal-first shell for AI agents: task-bound tabs, briefs, cross-agent reviews, the Studio, delivery gates. Writes tasks, phases, verdicts and artifacts to the platform.

You are here

The platform · twindem.ai

Admin · Board · Docs

One account and one memory. The Board is the source of truth; Docs hold architecture and approved designs; agents reach both through the official Twindem MCP tools — no API keys to manage.

See the Board

Communicate

Twindem Pulse

Native mail & calendar with the Vio assistant: emails become board tasks, digests brief your day, and project context follows you into every conversation.

Meet Pulse

What Team does

Everything below is built and running today.

Context-first agents

Agents that start briefed

Claude Code and Codex open already knowing the task, the history and the approved designs — via the official Twindem MCP tools and the twd-* skills.

How agents get context

The board is the source of truth

Work that records itself

Capture a task with a brief, and the board tracks reality: phases move when work starts, whoever works it owns it, every agent turn lands on the story.

See the delivery flow

Reviews with teeth

Cross-agent review, typed verdicts

A second agent — even from the other vendor — reads the record, reviews the work and posts a structured verdict. Open findings block Done at the server.

How gating works

The Studio

Designs & diagrams, live

UI mockups, C4 architecture and sequence diagrams proposed by agents, previewed live beside the terminal, approved into Docs as must-follow artifacts.

Inside the Studio

Peripheral vision

Know without watching

Ambient chips per tab, a two-second HUD, small top-right notifications with per-class sounds, a Needs-me queue, session digests and focus mode.

The awareness layer

Why it matters

Ship with agents. Keep the team.

AI agents multiply output — and, unmanaged, they multiply chaos. Team keeps the human rituals that make software trustworthy: a brief before work, a record of decisions, a second reviewer, and a gate before Done. The speed of agents, the discipline of a good team.

A team working together around a table full of laptops

Private preview

Bring your agents. Keep the proof.

Twindem Team is in private preview on macOS. Tell us about your team and we'll get you in.

Request early access