Deco CMS

Open-source · TypeScript-first · Deploy anywhere

Open-source control plane for your AI agents.

Docs · Discord · decocms.com/studio

TL;DR: Hire agents. Connect tools. Organize them into projects with a UI that fits the job. Track every token and dollar. Self-host or use the cloud.

Studio is where you hire agents, connect tools, and organize them into projects that actually do things.

Agents come with real skills and battle-tested prompts. Connections give them access to your tools — GitHub, Slack, Postgres, OpenRouter, anything that speaks MCP — set up through a web UI with one-click OAuth. Projects bring agents and connections together around a goal: each project gets its own sidebar and UI, shaped by what's inside it.

Everything is tracked — tokens, costs, errors, latency — per connection, per agent. Install locally and it stays private. Sync to the cloud for remote access, team roles, and shared billing.

┌─────────────────────────────────────────────────────────────────┐
│                           Clients                               │
│         Cursor · Claude · VS Code · Custom Agents               │
└───────────────────────────┬─────────────────────────────────────┘
                            │
                            ▼
┌─────────────────────────────────────────────────────────────────┐
│                         DECO CMS                              │
│  Agents · Connections · Projects · Observability · Token Vault  │
└───────────────────────────┬─────────────────────────────────────┘
                            │
                            ▼
┌─────────────────────────────────────────────────────────────────┐
│                     Tools & MCP Servers                          │
│      GitHub · Slack · Postgres · OpenRouter · Your APIs         │
└─────────────────────────────────────────────────────────────────┘

bunx decocms

Or clone and run from source:

git clone https://github.com/decocms/studio.git
bun install
bun run dev

runs at http://localhost:3000 (client) + API server

Browse and hire specialized AI agents from the built-in store. Each agent knows how to use specific tools and comes with battle-tested prompts. Agents compose — they can call each other — and every action is tracked with cost attribution.

Set up MCP connections through a web UI with one-click OAuth. No JSON configs. Monitor latency, errors, and costs per connection. Share tools across your team without sharing credentials.

As tool surfaces grow, Studio exposes Virtual MCPs — one endpoint, different strategies for which tools to surface:

• Full-context: expose everything (simple, deterministic, good for small toolsets)
• Smart selection: narrow the toolset before execution
• Code execution: load tools on demand in a sandbox

Projects bring agents and connections together around a goal. The project's UI adapts to what's inside — add a content agent and a CMS connection, the sidebar shows content management; add an analytics agent and a database, it shows dashboards and queries. The UI you see is the UI that's relevant for operating that project.

You can also define outcomes declaratively and let Studio work backward to derive milestones, assign agents, and verify results.

Token spend per connection — OpenRouter, Perplexity, Firecrawl, all of it. Latency, errors, bottlenecks. One dashboard.

Type-safe, audited, observable, callable via MCP.

import { z } from "zod";
import { defineTool } from "~/core/define-tool";

export const CONNECTION_CREATE = defineTool({
  name: "CONNECTION_CREATE",
  description: "Create a new MCP connection",
  inputSchema: z.object({
    name: z.string(),
    connection: z.object({
      type: z.enum(["HTTP", "SSE", "WebSocket"]),
      url: z.string().url(),
      token: z.string().optional(),
    }),
  }),
  outputSchema: z.object({
    id: z.string(),
    scope: z.enum(["workspace", "project"]),
  }),
  handler: async (input, ctx) => {
    await ctx.access.check();
    const conn = await ctx.storage.connections.create({
      projectId: ctx.project?.id ?? null,
      ...input,
      createdById: ctx.auth.user!.id,
    });
    return { id: conn.id, scope: conn.projectId ? "project" : "workspace" };
  },
});

Every tool call gets input/output validation, access control, audit logging, and OpenTelemetry traces automatically.

├── apps/
│   ├── mesh/                # Full-stack Deco CMS (Hono API + Vite/React)
│   │   ├── src/
│   │   │   ├── api/         # Hono HTTP + MCP proxy routes
│   │   │   ├── auth/        # Better Auth (OAuth + API keys)
│   │   │   ├── core/        # MeshContext, AccessControl, defineTool
│   │   │   ├── tools/       # Built-in MCP management tools
│   │   │   ├── storage/     # Kysely DB adapters
│   │   │   ├── event-bus/   # Pub/sub event delivery system
│   │   │   ├── encryption/  # Token vault & credential management
│   │   │   ├── observability/  # OpenTelemetry tracing & metrics
│   │   │   └── web/         # React 19 admin UI
│   │   └── migrations/      # Kysely database migrations
│   └── docs/                # Astro documentation site
│
└── packages/
    ├── bindings/            # Core MCP bindings and connection abstractions
    ├── runtime/             # MCP proxy, OAuth, and runtime utilities
    ├── ui/                  # Shared React components (shadcn-based)
    ├── cli/                 # CLI tooling (deco commands)
    ├── create-deco/         # Project scaffolding (npm create deco)
    └── vite-plugin-deco/    # Vite plugin for Deco projects

bun install          # Install dependencies
bun run dev          # Run dev server (client + API)
bun test             # Run tests
bun run check        # Type check
bun run lint         # Lint
bun run fmt          # Format

bun run dev:client     # Vite dev server (port 4000)
bun run dev:server     # Hono server with hot reload
bun run migrate        # Run database migrations

dev:worktree routes http://<WORKTREE_SLUG>.localhost via Caddy — useful for running multiple workspaces without port conflicts.

# One-time setup
brew install caddy && caddy start

# Start
WORKTREE_SLUG=my-feature bun run dev:worktree

# Conductor adapter (sets WORKTREE_SLUG from CONDUCTOR_WORKSPACE_NAME)
bun run dev:conductor

# Docker (embedded PostgreSQL)
docker compose -f deploy/docker-compose.yml up

# Docker (PostgreSQL)
docker compose -f deploy/docker-compose.postgres.yml up

# Bun
bun run build:client && bun run build:server && bun run start

# Kubernetes (Helm)
helm install deco-studio oci://ghcr.io/decocms/chart-deco-studio --version <version> -n deco-studio --create-namespace

No vendor lock-in. Runs on Docker, Kubernetes, AWS, GCP, or local runtimes.

• Agent marketplace — discover, hire, and compose agents
• Declarative planning engine
• Cost analytics and spend caps
• Remote access from any browser
• Live tracing debugger
• Workflow orchestration with guardrails

Sustainable Use License (SUL) — see LICENSE.md.

• Free to self-host for internal use
• Free for client projects (agencies, SIs)
• Commercial license required for SaaS or revenue-generating production systems

Questions? builders@decocms.com

bun run fmt      # Format
bun run lint     # Lint
bun test         # Test

See AGENTS.md for coding guidelines.