simstudioai/simView GitHub Homepage for Latest Official Releases

An open-source AI agent workflow builder, providing a lightweight visual interface for quickly building and deploying LLM applications that connect various tools.

Apache-2.0TypeScriptsimsimstudioai 14.0k Last Updated: September 04, 2025

Sim Studio - Open-Source AI Agent Workflow Builder

Project Overview

Sim Studio is a powerful, user-friendly platform that allows developers and agents to build, test, and optimize AI agent workflows. It is an open-source AI agent workflow builder, offering a lightweight and intuitive interface to quickly construct and deploy Large Language Models (LLMs) connected to various tools.

Key Features

🎨 Visual Workflow Design

Drag-and-drop Interface: Provides a Figma-like environment to build and test AI agents, allowing users to visually construct complex workflows.
Intuitive Canvas: Users can design automation tasks like drawing a flowchart.
What You See Is What You Get (WYSIWYG): Real-time preview of workflow execution.

🔧 Rich Tool Integration

The platform can easily connect agents to various services such as Gmail, Slack, Pinecone, Supabase, and more. Supported tools include:

Communication Tools: Gmail, Slack, Microsoft Teams, Telegram, WhatsApp
Data Storage: Supabase, Pinecone, Qdrant, S3
Productivity Tools: Notion, Google Docs, Google Sheets, Airtable
Development Tools: GitHub, Jira, Linear
AI Services: Hugging Face, ElevenLabs, Image Generator
Search Tools: Google Search, Perplexity, Tavily, Exa

🤖 Multi-Model Support

Supports multiple LLM providers:

OpenAI Models: GPT-4o, o1, o3, o4-mini, gpt-4.1
Anthropic Models: Claude 3.7 Sonnet
Google Models: Gemini 2.5 Pro, Gemini 2.0 Flash
Other Providers: Groq, Cerebras, xAI, DeepSeek
Local Deployment: Supports local models via Ollama

🚀 Flexible Deployment Options

Completed AI agent workflows can be deployed as APIs or chat interfaces:

Manual Trigger: Manually execute workflows.
API Deployment: Deploy workflows as RESTful APIs.
Scheduled Execution: Set agents to run automatically at specific times or intervals.
Webhook Trigger: Triggered by incoming webhooks (e.g., Slack messages).
Chat Instance: Deploy as a standalone chat application.

📊 Performance Monitoring

Monitor workflow costs and execution time for each step using logs:

Real-time performance tracking
Cost analysis
Execution time statistics
Error logging

Technical Architecture

Frontend Tech Stack

Framework: Next.js (App Router)
Runtime: Bun
UI Components: Shadcn, Tailwind CSS
State Management: Zustand
Flow Editor: ReactFlow
Real-time Communication: Socket.io

Backend Tech Stack

Database: PostgreSQL with Drizzle ORM
Vector Extension: pgvector (for AI embeddings and semantic search)
Authentication: Better Auth
Background Tasks: Trigger.dev
Documentation System: Fumadocs
Project Management: Turborepo (monorepo)

AI Functionality

Sim uses vector embeddings for AI functionalities such as knowledge bases and semantic search, which requires the pgvector PostgreSQL extension.

Installation and Usage

Quick Start (NPM)

npx simstudio

After launching, visit http://localhost:3000/

Docker

# Clone the repository
git clone https://github.com/simstudioai/sim.git
cd sim

# Start Sim
docker compose -f docker-compose.prod.yml up -d

Local Model Support

# GPU support
docker compose --profile local-gpu -f docker-compose.ollama.yml up -d

# CPU support
docker compose --profile local-cpu -f docker-compose.ollama.yml up -d

Development Environment Setup

# Install dependencies
git clone https://github.com/simstudioai/sim.git
cd sim
bun install

# Set up PostgreSQL database
docker run --name simstudio-db \
  -e POSTGRES_PASSWORD=your_password \
  -e POSTGRES_DB=simstudio \
  -p 5432:5432 -d \
  pgvector/pgvector:pg17

# Configure environment variables
cd apps/sim
cp .env.example .env

# Database migration
bunx drizzle-kit migrate

# Start development server
bun run dev:full

Core Functional Modules

Agent Block

The Agent block serves as the interface between the workflow and the Large Language Model (LLM). It executes inference requests to various AI providers, processes natural language input based on defined instructions, and generates structured or unstructured output for downstream use.

Tool System

Tools extend the agent's capabilities by integrating external APIs and services. The tool system supports function calling, allowing agents to perform actions beyond text generation.

Structured Output

Response format parameters enforce structured output generation through JSON Schema validation.

Use Cases

Business Automation

Customer service automation
Data analysis and report generation
Email and message processing
Content creation and management

Research and Development

Information gathering and organization
Code review and management
Project management automation
Knowledge base construction

Integrated Applications

CRM system integration
Social media management
E-commerce platform automation
Database operation automation

Key Advantages

Low-code/No-code: The drag-and-drop approach eliminates common barriers in agent development, such as connecting data sources, requiring less manual work.
Open Source: Fully open-source, community-driven development.
Multi-Model Support: Supports mainstream AI providers and local models.
Enterprise-grade: Supports self-hosting, ensuring data security and control.
Rich Ecosystem: Integrates with a large number of third-party services.
Performance Monitoring: Built-in performance analysis and cost tracking.

Project Team

Sim Studio was founded in 2025 by Emir Karabeg and Waleed Latif. The team consists of 2 employees and is headquartered in San Francisco, California, USA.

Emir Karabeg (Co-founder & CEO): Studied Data Science and Cognitive Science at UC Berkeley, previously conducted NLP machine translation research, and built the WorkNinja AI learning platform.
Waleed Latif (Co-founder & CTO): Studied Computer Science and Cognitive Science at UC Berkeley, previously built core backend infrastructure at Amazon's Ring division.

Open Source License

This project is released under the Apache License 2.0. Community contributions are welcome.