DISCOVER THE FUTURE OF AI AGENTSarrow_forward

deco Studio

calendar_todayAdded Apr 22, 2026
categoryAgent & Tooling
codeOpen Source
TypeScriptNode.jsWorkflow AutomationDockerMulti-Agent SystemModel Context ProtocolAI AgentsWeb ApplicationAgent & ToolingAutomation, Workflow & RPAProtocol, API & Integration

Open-source control plane for AI agents — connect tools via MCP, hire and compose agents, track every token and dollar with full observability.

deco Studio is an open-source AI agent control plane built around the Model Context Protocol (MCP). It provides an Agent Store (browse, hire, and compose agents), one-click OAuth connections for 50+ tools, Token Vault for centralized credential management, and Virtual MCPs with three modes — Full-context, Smart selection, and Code execution. On observability, it tracks token consumption, latency percentiles, and error rates per connection and agent, with native OpenTelemetry integration. Projects are organized via a declarative planning engine that defines outcomes and lets Studio work backward to derive milestones and assign agents (current implementation depth unconfirmed, listed as future item in Roadmap). Multi-tenancy is powered by Better Auth with RBAC, isolating configurations, credentials, and audit logs at the Workspace/Project level. A built-in event bus supports inter-connection Pub/Sub with scheduled delivery (At-least-once). The stack runs on Bun with Hono backend, React 19 frontend, Kysely database layer, and tools defined declaratively via defineTool + Zod Schema with automatic validation, access control, audit logging, and tracing. Supports one-click install (bunx decocms), Docker, and Kubernetes (Helm) deployment with full self-hosting capability. Currently in Public Beta under the Sustainable Use License (SUL) — free for internal use and client projects, commercial licensing required for SaaS.

Installation & Deployment#

bunx decocms

From source:

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

Docker (embedded PostgreSQL):

docker compose -f deploy/docker-compose.yml up

Docker (external PostgreSQL):

docker compose -f deploy/docker-compose.postgres.yml up

Kubernetes (Helm):

helm install deco-studio oci://ghcr.io/decocms/chart-deco-studio --version <version> -n deco-studio --create-namespace

Production build:

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

Unconfirmed Information#

  • Declarative planning engine is listed as a future Roadmap item; current implementation depth is unconfirmed
  • Official website mentions SQLite, but project structure shows Kysely → PostgreSQL — a contradiction that needs clarification
  • Actual number and types of hireable agents in the Agent Store are not publicly documented
  • Specific SUL commercial licensing pricing and terms require reviewing LICENSE.md
  • The full product matrix relationship between the decocms organization name and deco Studio product name is unclear

Related Projects

View All arrow_forward

STAY UPDATED

Get the latest AI tools and trends delivered straight to your inbox. No spam, just intelligence.

rocket_launch