Automation API
The Automation API (Workspace REST API) lets agents, CI, and internal tools operate a Vibe workspace without clicking through Admin.
Open Workspace Settings → API to create keys and copy the live API documentation and base URL (published on your workspace apps publisher identity, typically *.diode.link on the publisher port).
Auth
- Create a key under Workspace Settings → API (
Create API key…) - Send
Authorization: Bearer dv_ws_…on every authenticated request - Keys begin with
dv_ws_. Workspace-wide plaintext is shown once at creation — store it securely - Choose all projects or one project, plus which endpoint scopes are enabled — both are immutable after create (mint a new key to change)
Project-limited keys can only touch that project (and workspace/info); they cannot create projects. Full request/response shapes live at the docs URL in Workspace Settings.
Endpoints (summary)
| Endpoint | Purpose |
|---|---|
POST /api/v1/project/list |
List projects in this workspace |
POST /api/v1/project/info |
Project details, including Nexus generation status when in progress |
POST /api/v1/project/health |
Reconciled runtime health (Development Mode + project/health scope) |
POST /api/v1/project/logs |
Bounded App / Diode runtime logs (Development Mode + project/logs scope) |
POST /api/v1/project/create |
Create a project shell (name, type, description, publish ports/scope) — does not run Nexus |
POST /api/v1/project/update |
Update publish settings (external port, local port, scope, allowlist) |
POST /api/v1/project/generate |
Start Nexus generation; poll project/info for generation.job progress |
POST /api/v1/project/delete |
Delete a project by project_id |
POST /api/v1/workspace/info |
Workspace balance, usage, and billing plan |
Deploy uploads still use the HTTPS /deploy path from each project’s Agent Instructions — the Automation API is for workspace automation around create / update / generate / health / logs.
Why it matters
This is how Vibe stays agent-first: Cursor, Claude, or a CI bot can create the right project type, adjust publish settings, kick off Nexus, and verify health — the same ladder described in Agent-driven deploy.