StellaryStellaryBeta
FeaturesHow It WorksPlansBlog
Overview
Concepts & architecture
Getting Started
Workspace, project, context, and tokens
API Reference
Backend routes, auth, and models
MCP Integration
MCP clients, agents, and workspace tools
FAQ
Sign inTry for free
FeaturesHow It WorksPlansBlog
Documentation
Overview
Concepts & architecture
Getting Started
Workspace, project, context, and tokens
API Reference
Backend routes, auth, and models
MCP Integration
MCP clients, agents, and workspace tools
?
FAQ
Sign inTry for free
StellaryStellary

The multi-agent command center for teams that ship.

Product

  • Features
  • How It Works
  • Plans
  • Blog
  • FAQ

Developers

  • Documentation
  • API Reference
  • MCP Integration
  • Getting Started

Company

  • About
  • Product ambitions
  • Editorial policy
  • How we compare tools
  • Legal Notice
  • Terms of Service
  • Privacy Policy
  • Cookie Policy
  • DPA

© 2026 Stellary. All rights reserved.

Legal NoticeTerms of ServicePrivacy PolicyCookie PolicyDPA
Overview
Guide
  • User Guide
  • Board & Cards
  • Knowledge Base
  • Cockpit & Command Center
  • AI Project Wizard
  • AI Agents & MCP
  • Automations
  • Team & Collaboration
Developers
  • Getting Started
  • API Reference
  • MCP Integration

Getting Started

Start from a fresh account and end with a workspace that can run delivery, documents, pilotage, and AI integrations. This guide follows the route model and objects that exist in Stellary today.

If you already know the product, jump straight to the API Reference or the MCP guide.

1Create account2Create project3Structure delivery4Add documents5Invite team6Pilotage7Connect agents8Tokens & automations

1 Create your account and first workspace

Go to https://app.stellary.co/register and create an account, or sign in if a workspace has already been shared with you.

On self-signup, Stellary automatically provisions:

  • A personal organization
  • A personal workspace inside that organization

The URL of that workspace gives you the :orgSlug and :wsSlug used throughout the REST and MCP docs.

2 Create a project or start with a draft

If you already know the delivery shape, create a project directly. If you want Stellary to help turn raw context into a structured starting point, create a project draft first.

POST /orgs/:orgSlug/workspaces/:wsSlug/projects
{
"name": "Checkout stabilization",
"description": "Reduce payment failures and improve rollback safety",
"deadline": "2026-06-01T00:00:00.000Z"
}
POST /orgs/:orgSlug/workspaces/:wsSlug/project-drafts
{
"name": "Checkout stabilization",
"description": "Give the agent the business context first",
"contextDocuments": ["Existing incident review", "Support pain points"],
"teamMembers": [
{ "name": "Alex", "role": "Engineering lead" }
]
}

3 Structure delivery with scopes and cards

The core delivery model is project -> scope -> card. In the UI, the same project can be explored through board, list, calendar, and stats views, while roadmap and timeline data also live on the project delivery surface.

POST /orgs/:orgSlug/workspaces/:wsSlug/projects/:projectId/scopes
{
"name": "Payment resilience",
"type": "kanban"
}
POST /orgs/:orgSlug/workspaces/:wsSlug/projects/:projectId/cards
{
"title": "Ship rollback-safe payment webhook",
"description": "Add retries, alerts, and idempotency checks",
"priority": "high",
"dueDate": "2026-04-20T12:00:00.000Z"
}

From there you can add comments, attachments, labels, relations, assignees, dates, and custom fields as the delivery model becomes more specific.

4 Add documents and shared context

Documents are workspace-scoped resources that can belong to an organization, workspace, project, or scope. They accept both markdown and tiptap_json content.

Supported document types in the backend today are:

  • document, spreadsheet, spec, brief
  • adr, note, reference, template
POST /orgs/:orgSlug/workspaces/:wsSlug/documents
{
"title": "Payment retry policy",
"content": "# Payment retry policy",
"contentFormat": "markdown",
"docType": "spec",
"contextType": "project",
"contextId": "prj_123",
"tags": ["payments", "resilience"]
}

When a document should inform execution directly, link it to a card through /projects/:projectId/cards/:cardId/documents/:documentId/link.

5 Invite the team and define access

Start at workspace level unless you have a strong reason to manage access elsewhere. That is where projects, documents, agents, pipelines, and most day-to-day operations converge.

POST /orgs/:orgSlug/workspaces/:wsSlug/members/invite
{
"email": "teammate@company.com",
"role": "member"
}
SurfaceRoles
Organizationowner, admin, member
Workspaceowner, admin, member
Projectowner, member, agent
Pilotageviewer, reviewer, owner

6 Turn on pilotage when delivery starts moving

Pilotage is the steering layer for a workspace. Use it to concentrate signals from active work into priorities, missions, decisions, dashboard views, and proposed actions.

GET /orgs/:orgSlug/workspaces/:wsSlug/pilotage/dashboard
POST /orgs/:orgSlug/workspaces/:wsSlug/pilotage/actions/propose

This is also the review surface where AI-generated changes can be examined before they are accepted in more controlled setups.

7 Connect agents and MCP clients

Stellary distinguishes between agent configuration and agent runtime. Create a workspace agent first, then run it directly or attach it to card missions.

POST /orgs/:orgSlug/workspaces/:wsSlug/agents
{
"name": "Delivery reviewer",
"slug": "delivery-reviewer",
"description": "Reviews risky work before apply",
"autonomyMode": "approval"
}

Runtime flows then use endpoints such as:

POST /orgs/:orgSlug/workspaces/:wsSlug/agent-runtime/run
POST /orgs/:orgSlug/workspaces/:wsSlug/agent-runtime/card-mission
GET /mcp
POST /mcp

The current autonomy modes exposed in the product surface are supervised, approval, and autonomous.

// .cursor/mcp.json
{
"mcpServers": {
"stellary": {
"url": "https://app.stellary.co/mcp",
"headers": {
"Authorization": "Bearer YOUR_TOKEN_HERE"
}
}
}
}

Continue with the MCP Integration guide for client-by-client setup details.

8 Create API tokens and project automations

Personal access tokens work for REST and MCP. They are the safest default for CI, scripts, and external AI clients.

POST /api-tokens
{
"name": "CI pipeline",
"scopes": ["projects:read", "projects:write"],
"expiresAt": "2027-01-01T00:00:00.000Z"
}

The available token scopes in the shared types are:

  • projects:read, projects:write
  • pilotage:read, pilotage:write
  • notifications:read
  • account:read, account:write

Project automations sit under the delivery surface:

POST /orgs/:orgSlug/workspaces/:wsSlug/projects/:projectId/automations

Supported trigger identifiers currently include:

  • task_moved, task_created, checklist_complete
  • due_date_reached, task_status_changed, task_assigned
  • task_added_to_board
API ReferenceMCP IntegrationBoard & Cards Guide

If you are evaluating the category before diving deeper, continue with AI project management software in 2026 or Project management with AI agents.