Gives AI clients a single governed MCP endpoint for agents, tools, models, and per-call cost tracking.
Also known as decocms/studio, deco-cli, decocms
- Stars
- 418
- Tools
- —
- Downloads, 30 days
- 85,291
- Latest version
- —
- Last commit
- Nov 14, 2026
- License
- MIT
What is deco Studio MCP?
deco Studio gives AI clients a single governed MCP endpoint for agents, tools, models, and per-call cost tracking.
Summary generated from the project's README
About deco Studio MCP server
Providing a centralized control plane for organizational AI deployment is what studio does. It acts as a single governed MCP endpoint that manages agents, tools, and models while tracking per-call costs. Organizations can use it to package context and policy into publishable agents that are callable from clients like Cursor, Claude Desktop, or VS Code.
Users can connect private systems such as GitHub, Slack, or Postgres through an encrypted vault, allowing teams to share MCP capabilities without sharing actual credentials. The platform supports interchangeable model layers, enabling connections to OpenRouter or direct providers like Anthropic, OpenAI, and Google. It can be deployed locally using Docker Compose with PostgreSQL and NATS, or used via a cloud service for remote team access.
Administrators maintain oversight through observability dashboards that trace latency, errors, and tokens. The system provides enterprise-grade management features including SSO, role-based access control, and audit logs to ensure secure and accountable AI usage across the entire company.
What you can do with it
- 1Securely share internal toolsA developer can grant an agent access to a private database without exposing the underlying credentials.
- 2Monitor AI operational costsAn administrator can track exactly how much money is being spent on different models and agents.
- 3Standardize agent configurationsTeams can create specific agents with predefined instructions and tools to be used by the whole organization.
- 4Manage multi-model routingAn engineer can switch between different AI providers for different tasks through a single interface.
Written from the project's README. Every use case is checked against a quote from it.
How to install deco Studio MCP
Formats verified 10 Nov 2026Add to CursorAdd to GooseInstall deco Studio MCP in Claude Code
$ claude mcp add studio -- npx -y deco-cliInstall deco Studio MCP in Cursor
~/.cursor/mcp.json
{
"mcpServers": {
"studio": {
"command": "npx",
"args": [
"-y",
"deco-cli"
]
}
}
}Install deco Studio MCP in VS Code
.vscode/mcp.json
{
"servers": {
"studio": {
"type": "stdio",
"command": "npx",
"args": [
"-y",
"deco-cli"
]
}
}
}Install deco Studio MCP in VS Code CLI
$ code --add-mcp '{"name":"studio","type":"stdio","command":"npx","args":["-y","deco-cli"]}'Install deco Studio MCP in Claude Desktop
claude_desktop_config.json
{
"mcpServers": {
"studio": {
"command": "npx",
"args": [
"-y",
"deco-cli"
]
}
}
}Install deco Studio MCP in Codex
$ codex mcp add studio -- npx -y deco-cliInstall deco Studio MCP in Codex config.toml
~/.codex/config.toml
[mcp_servers.studio] command = "npx" args = ["-y", "deco-cli"]
Install deco Studio MCP in Gemini CLI
$ gemini mcp add -s user studio npx -- -y deco-cliInstall deco Studio MCP in Gemini settings.json
~/.gemini/settings.json
{
"mcpServers": {
"studio": {
"command": "npx",
"args": [
"-y",
"deco-cli"
]
}
}
}Automated checks
Ran Nov 12, 2026Full report →Passed our automated checks. That is not a security audit: read the code and give the server the narrowest access it needs. How checks work
- Source repository linked and reachablePass
- Open source license (OSI)Pass
- Commit or release in the last yearPass
- Package name not close to a popular packagePass
- Remote endpoint uses HTTPSNot checked
- Destructive tools are labelledNot checked
- No hidden instructions to the modelPass
- No critical advisory in the package or its dependenciesNot checked
README
View on GitHubTL;DR: Your team needs a secure internal vibecoding platform. You just found it. Configure agents with team context. Connect private MCPs once — share capabilities, not credentials. Keep the model layer interchangeable. Roll out across the organization with SSO, RBAC, audit logs, and cost controls — all through one MCP endpoint. Local-first. Self-host or use the cloud.
What is deco Studio?
Studio packages the infrastructure behind an internal AI rollout: model routing, MCP authentication, agent configuration, SSO, RBAC, audit logs, and usage accounting. Your teams get chat. You keep control.
Under the hood it's one control plane for your AI agents — one MCP endpoint for all your agents, tools, and models. Agents package context, tools, and policy into something you publish to the organization. Connections give them governed access to your systems — GitHub, Slack, Postgres, Sentry, anything that speaks MCP — with tokens stored in an encrypted vault. Models stay interchangeable: OpenRouter or direct providers, chosen per agent and per tool.
Start with one team. Standardize approved models, tools, and context. Expand across the organization without copying secrets or rebuilding the platform. 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 STUDIO │
│ Agents · Connections · Models · Vault · Observability │
└───────────────────────────┬─────────────────────────────────────┘
│
▼
┌─────────────────────────────────────────────────────────────────┐
│ Tools & MCP Servers │
│ GitHub · Slack · Postgres · OpenRouter · Your APIs │
└─────────────────────────────────────────────────────────────────┘
Quick Start
Install Docker with Compose, then run from source:
git clone https://github.com/decocms/studio.git
cd studio
docker compose up --build --watch
Open http://localhost:4000. Compose starts PostgreSQL, NATS, object storage, migrations, and the hot-reloading API and web app. See CONTRIBUTING.md for contributor setup and worktrees.
What you get
Agents
Package context, tools, and policy into an agent. Define instructions, add skills and files, grant approved MCP access, choose a model policy, then publish the agent to the organization. Each agent is its own MCP endpoint — callable from Cursor, Claude Desktop, your own code, or another agent. Agents compose, and every action is tracked with cost attribution.
Connections
Connect private systems once, securely. Register MCP servers at the organization level through a web UI with one-click OAuth — no JSON configs. Tokens live in the encrypted vault, and you grant tool-level access by organization, role, or agent. Share MCP capabilities — not 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
Models
Keep the AI layer interchangeable. Use OpenRouter or connect Anthropic, OpenAI, Google, or any compatible provider directly — the best model for each agent and tool, behind one router. For coding work, engineers can link their own Claude Code or Codex session and use the subscription already authenticated on their machine.
Projects
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.
Observability
Account for every model and tool call. Trace the user, agent, model, tools, latency, errors, tokens, and cost for every thread. Break usage down by agent, connection, organization, or teammate — one dashboard.
From your desktop to your org
| Local | Docker Compose on your laptop, including PostgreSQL, NATS, and object storage. |
| Cloud | Log in to studio.decocms.com. Control local projects from any browser. |
| Team | Invite people. SSO and role-based access. Shared connections. Cost attribution. |
| Enterprise | Self-hosted. Organization isolation, tool-scoped API keys, audit logs. Your infra, your rules. |
Core Capabilities
| Capability | What it does |
|---|---|
| Agents | Package context, tools, and policy into publishable agents with cost attribution |
| Connections | Route MCP traffic through one governed endpoint with auth, proxy, and encrypted token vault |
| Models | Interchangeable AI layer — OpenRouter or direct providers, model policy per agent |
| Projects | Organize agents and connections around goals with an adaptive UI |
| Virtual MCPs | Compose and expose governed toolsets as new MCP endpoints |
| Observability | Traces, costs, errors, and latency per user, agent, and connection — one dashboard |
| Access Control | SSO + RBAC via Better Auth — OAuth 2.1 and tool-scoped API keys per workspace/project |
| Multi-tenancy | Organization/project isolation for config, credentials, policies, and audit logs |
| Bindings | Capability contracts so tools target interfaces, not specific implementations |
| Store | Discover and install agents, tools, and templates |
Define Tools
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.
Documentation
Studio documentation lives in docs-tanstack, alongside the Storefront documentation. Update public product guides and architecture there. Workspace READMEs and engineering plans stay here with the implementation.
Project Structure
Applications
| Workspace | Purpose |
|---|---|
apps/api |
Hono API, authentication, tools, storage, migrations, and the deco CLI |
apps/web |
Vite and React 19 administration interface |
Packages
| Workspace | Purpose |
|---|---|
packages/bindings |
Runtime-validated MCP capability contracts |
packages/create-deco |
npm create deco project scaffolding CLI |
packages/e2e |
Black-box Playwright suite for Studio |
packages/harness-runner |
In-sandbox coding-agent harness process |
packages/mcp-utils |
MCP proxy, gateway, aggregation, and sandbox primitives |
packages/runtime |
Runtime helpers for MCP servers, OAuth, tools, and triggers |
packages/sandbox |
Agent sandbox lifecycle, daemon, dispatch, and proxy implementation |
packages/sandbox-controller |
Sandbox provider bundled for a remote host such as the control plane |
packages/shared |
Private isomorphic contracts, SDK utilities, and async primitives |
packages/typegen |
Typed client generator for Studio Virtual MCPs |
packages/ui |
Internal React design system |
Development
bun install # Install dependencies
bun run dev # Docker Compose Watch: dependencies, migrations, API + web
bun test # Run tests
bun run check # Type check
bun run lint # Lint
bun run fmt # Format
App-specific commands
bun run --cwd=apps/web dev # Vite dev server (port 4000)
bun run --cwd=apps/api dev:server # Hono server with hot reload
bun run --cwd=apps/api migrate # Run database migrations
For parallel worktrees, Compose lifecycle commands, and frontend-only work, see CONTRIBUTING.md.
Deploy Anywhere
# Docker Compose: released Studio + PostgreSQL + NATS + object storage
docker compose -f selfhost/examples/docker-compose/compose.yaml up -d
# Kubernetes (Helm)
helm install deco-studio oci://ghcr.io/decocms/chart-deco-studio --version <version> -n deco-studio --create-namespace
Run Studio with Docker Compose or Kubernetes on your own infrastructure.
What you need to run it
| Tier | Footprint |
|---|---|
| Laptop | Docker with Compose; the development stack includes its dependencies. |
| Docker | Published Studio image with configured PostgreSQL, NATS, and object storage. |
| Production (Helm) | PostgreSQL you bring, plus NATS for streaming and cross-pod coordination, ClickHouse + OTel Collector (traces and analytics), and the sandbox operator (isolated agent environments on Kubernetes). Your identity provider, your model keys, your storage. |
Production topology
graph TB
clients["MCP clients — Cursor · Claude · VS Code · your code"]
clients -->|"one MCP endpoint · SSO · RBAC · audit"| api
subgraph k8s ["Kubernetes (Helm)"]
api["Studio API + Admin UI"]
api --> sandbox["Agent sandboxes<br/>(sandbox-operator)"]
api <-->|"streams and broadcasts"| nats["NATS"]
api -->|"traces · costs"| otel["OTel Collector"]
nats <-->|"streams and broadcasts"| worker["Workers<br/>DBOS runs"]
otel --> ch[("ClickHouse")]
end
pg[("PostgreSQL")]
api --> pg
worker <-->|"DBOS queues and journal"| pg
subgraph upstream ["Models & tools"]
models["Anthropic · OpenAI<br/>OpenRouter · Ollama"]
mcps["GitHub · Slack · Postgres<br/>your MCP servers"]
end
api -->|"model routing · vaulted credentials"| upstream
Every box is optional except Studio and PostgreSQL — start small, turn on the rest as the rollout grows.
Tech Stack
| Layer | Tech |
|---|---|
| Runtime | Bun / Node |
| Language | TypeScript + Zod |
| Framework | Hono (API) + Vite + React 19 |
| Database | Kysely → PostgreSQL |
| Auth | Better Auth (OAuth 2.1 + API keys) |
| Observability | OpenTelemetry |
| UI | React 19 + Tailwind v4 + shadcn |
| Protocol | Model Context Protocol (MCP) |
Roadmap
- 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
License
MIT — see LICENSE.md.
Questions? builders@decocms.com
Contributing
bun run fmt # Format
bun run lint # Lint
bun test # Test
See AGENTS.md for coding guidelines, CONTRIBUTING.md for the contribution workflow, and TESTING.md for the testing rules.
Good first steps: open an issue, pick one labeled good first issue, or ask in Discord.
Contributors
Studio is built in the open. Thanks to everyone who has shipped a commit.
Maintainers — see MAINTAINERS.txt:
@guitavano · @viktormarinho · @mcandeia · @pedrofrxncx · @JonasJesus42 · @aka-sacci-ccr · @vibegui · @tlgimenes · @vibe-dex · @nicacioliveira
deco Studio MCP questions
How do I add deco Studio MCP to Claude Code?
Run claude mcp add studio -- npx -y deco-cli in your terminal. Add --scope user to make it available in every project.
How do I add deco Studio MCP to Cursor?
Open ~/.cursor/mcp.json and add the server under mcpServers, using the Cursor config shown in the install section above.
How do I run deco Studio locally?
Clone the repository and start it with Docker Compose, then open localhost:4000. Compose brings up PostgreSQL, NATS, object storage, migrations, and the API and web app.
How does Studio store credentials for connected tools?
MCP server tokens are kept in an encrypted vault at the organization level, and access is granted per organization, role, or agent.
Can I use different AI model providers?
Yes, Studio routes models through OpenRouter or direct providers such as Anthropic, OpenAI, and Google, chosen per agent and tool.
What can I track about agent usage?
Every model and tool call is traced with user, agent, model, latency, errors, tokens, and cost, broken down by agent, connection, organization, or teammate.
Does Studio support SSO and role-based access?
Yes, it offers SSO and RBAC via Better Auth with OAuth 2.1 and tool-scoped API keys, plus organization isolation and audit logs for self-hosted deployments.