This document provides a high-level introduction to Sim Studio, an AI agent workflow platform for building and deploying automated agentic workforces. It covers the system's purpose, core capabilities, architectural patterns, and technology stack.
Sim Studio is an open-source platform designed to visually orchestrate AI agents, tools, and LLMs. Users build workflows by connecting "blocks" (logic units) into directed acyclic graphs (DAGs) on a visual canvas README.md44 These workflows can be executed instantly, deployed as APIs, or triggered via webhooks and schedules README.md46-47
The system is architected as a Next.js monorepo using the Bun runtime for high-performance execution package.json3-10 apps/sim/package.json7-8
The platform features a visual canvas that allows for drag-and-drop workflow construction README.md37-38 It supports real-time collaborative editing powered by a dedicated Socket.io server in apps/realtime apps/realtime/package.json102
Sim Studio integrates with every major LLM provider, including OpenAI, Anthropic, Google Gemini, and local providers like Ollama and vLLM README.md42-102 apps/sim/package.json39-157 Agents can utilize a vast catalog of built-in tools to perform complex tasks, including specialized capabilities like stagehand for web navigation apps/sim/package.json65
The Mothership serves as an AI command center where users can describe tasks in plain language. The workspace home view provides a centralized interface for task management, template prompts, and AI interaction driven by the use-chat hook README.md44-50
The platform includes built-in structured data storage via Tables README.md56 retrieval-augmented generation (RAG) capabilities via a Knowledge Base powered by PostgreSQL with pgvector README.md65-119 and a data Enrichment system for cascading provider lookups (e.g., email verification, company info) apps/sim/package.json154
The following diagram illustrates the high-level relationship between the user interface, the application logic, and the infrastructure components.
Sim Studio Architecture Map
Sources: apps/sim/package.json130-172 apps/realtime/package.json101-102 README.md19-21 apps/sim/next.config.ts1-9
Sim Studio leverages a modern, type-safe stack designed for scale and developer velocity.
| Layer | Technology | Key Package/Version |
|---|---|---|
| Runtime | Bun | ^1.3.13 package.json3 |
| Framework | Next.js | 16.2.11 apps/sim/package.json76 |
| Frontend | React | 19.2.4 apps/sim/package.json74 |
| Database | PostgreSQL | ^3.4.5 (driver) apps/realtime/package.json100 |
| ORM | Drizzle | ^0.45.2 apps/sim/package.json148 |
| Auth | Better Auth | 1.6.23 apps/sim/package.json133 |
| Real-time | Socket.io | ^4.8.1 apps/realtime/package.json102 |
| Workflow Engine | Trigger.dev | 4.4.3 apps/sim/package.json130 |
For a comprehensive breakdown, see Technology Stack.
The project is organized as a Turborepo monorepo package.json7-10 separating the core application from shared libraries and the documentation site.
Code Entity Space Bridge
Sources: package.json7-10 apps/sim/package.json105-115 apps/realtime/package.json87-97 apps/docs/package.json20-22 apps/sim/next.config.ts1-29 docker/app.Dockerfile89
Sim Studio implements enterprise-grade security patterns:
isolated-vm to prevent SSRF and unauthorized system access apps/sim/package.json166 apps/sim/next.config.ts130@better-auth/sso apps/sim/package.json63-133getMainCSPPolicy and getWorkflowExecutionCSPPolicy to protect against XSS and data exfiltration apps/sim/next.config.ts5-225bootstrap.js entrypoint to hydrate runtime secrets docker/app.Dockerfile89-143Sources: apps/sim/package.json133-166 apps/sim/next.config.ts5-225 docker/app.Dockerfile1-143
Refresh this wiki
This wiki was recently refreshed. Please wait 3 days to refresh again.
