Specialized employees
Product, architecture, development, QA, and senior review roles that hand off cleanly.
DDesk
AI employees that ship — with you at the gates.
The delivery desk for product teams: specialized AI roles run the pipeline in isolated sandboxes; humans keep AC, PR, and production control.
How it works
From first ticket to production gate — usually under fifteen minutes to see a live story.
- 01
Connect a repo
Create a project and paste your HTTPS GitHub URL so sandboxes can clone safely.
- 02
Start a story
Open Stories (or Auto-fix a ticket). Pick Fast or Full lane — Spec Kit planning runs before Development.
- 03
Watch the desk
Product Owner → Spec Kit plan → Developer → QA → Tech Lead (Full adds Senior + DevOps).
- 04
Approve the gates
You decide story AC, pull requests, and production. AI never skips a human gate.
Employee pipeline
A roster of roles — not a single agent. Amber steps are hard human gates.
Product
AC gate
Spec Kit plan
Developer
QA
Tech Lead
PR gate
Done / Full+
Built for serious delivery
Orchestration, sandboxes, engines, and gates in one workspace — not another coding chat.
Pluggable engines
Cursor, Claude, OpenAI-compatible APIs, or CLI tools — pick what fits each task.
Isolated sandboxes
Ephemeral Docker or local workspaces clone your repo, apply changes, then destroy.
Artifacts & audit
Every story leaves a trail: plans, tests, reviews, PRs, costs, and deployment records.
Health → Fix
Scans and site analytics open problems; approve Cursor Fix jobs with the same gates.
API for developers
Machine keys (aie_…) and Swagger at /api/docs for CI and your own tools.
Who it’s for
Teams that already try AI coding tools and need governed delivery next.
Product & eng leads
Want AI speed without silent merges or mystery deploys.
SaaS teams (5–40 eng)
Ship weekly and need a desk, not another autocomplete plugin.
Operators & founders
Need audit trails, cost visibility, and human approval as policy.
Open your desk
Free to start. Connect a repo, create a story, and decide the first gate.