# SIM using Ollama: 100% Local AI Agents Builder! (No Code)

Source: https://www.youtube.com/watch?v=jyWbD_Tp-EA
Recap page: https://rapidrecap.app/video/jyWbD_Tp-EA
Generated: 2025-09-04T07:32:30.751+00:00

---
## Quick Overview

This video demonstrates how to build and deploy AI agents locally using SIM, a no-code platform, and Ollama for running AI models. It covers cloning the SIM repository, configuring environment variables, setting up a PostgreSQL database with pgvector, running the application using Bun, and deploying it as a chat interface.

**Key Points:**
- The video guides users through setting up and running AI agents locally using the SIM platform and Ollama.
- It details cloning the SIM repository from GitHub and navigating its file structure.
- The tutorial includes configuring the .env file with necessary database credentials.
- Users learn to set up a PostgreSQL database with the pgvector extension using Docker.
- The process involves running the SIM application using the Bun package manager with the command 'bun run dev:full'.
- The video shows how to create and deploy a chat interface for the AI agent, allowing users to interact with it via a web browser.
- It demonstrates adding custom knowledge bases to the AI agents for enhanced functionality.

![Screenshot at 00:07: The video displays the SIM platform's visual interface, showing multiple agent workflows being constructed with 'Start', 'Agent 1', 'Agent 2', and 'Response' blocks, illustrating the process of building AI agent pipelines.](https://ss.rapidrecap.app/screens/jyWbD_Tp-EA/00-00-07.png)

**Context:** The video focuses on building and deploying AI agent workflows locally, emphasizing the use of the SIM platform, which is presented as an open-source, user-friendly environment for developers and agents. It highlights the ability to run AI models like Llama 3.2 locally using Ollama, without external API dependencies, and integrates with tools like PostgreSQL for data management.

## Detailed Analysis

This tutorial provides a comprehensive guide on setting up and running AI agents locally using the SIM platform and Ollama. It begins with cloning the SIM repository from GitHub, then guides the user through the local setup process. Key steps include configuring the .env file with database connection details, specifically for PostgreSQL with the pgvector extension, which is essential for AI agent functionality. The video demonstrates how to start the PostgreSQL database using Docker with a specific command. Subsequently, it shows how to run the SIM application using the Bun package manager by executing 'bun run dev:full' in the terminal, which initiates both the main Next.js application and the real-time socket server. The tutorial also covers adding custom knowledge bases to the AI agents, allowing them to access and utilize specific data. Finally, it walks through deploying the AI agent as a chat interface, which can be accessed via a personalized subdomain, enabling interactive use through a web browser. The process is presented as a no-code solution for building sophisticated AI agent workflows.

### Cloning and Setup

- Cloning SIM repository from GitHub
- Navigating project structure
- Configuring .env file with database credentials

### Database Setup

- Installing and running PostgreSQL with pgvector extension using Docker
- Verifying database status

### Running the Application

- Executing 'bun run dev:full' to start the SIM application
- Understanding the terminal output for running services

### Agent Creation

- Building agent workflows within the SIM platform
- Configuring system prompts, user prompts, and AI models (e.g., Llama 3.2)
- Adding tools to agents

### Knowledge Base Integration

- Creating and uploading custom knowledge bases for AI agents
- Utilizing knowledge bases for enhanced agent performance

### Deployment

- Deploying the AI agent as a chat interface
- Customizing subdomain and chat title
- Selecting chat output and access control options
- Testing the deployed chat interface

![Screenshot at 00:00: The video opens with a view of the SIM project's GitHub repository page, highlighting its open-source nature and the "Build/Deploy Agent Workflows" tagline.](https://ss.rapidrecap.app/screens/jyWbD_Tp-EA/00-00-00.png)
![Screenshot at 00:07: The SIM platform's visual workflow builder is shown, demonstrating the creation of AI agent pipelines with interconnected 'Start', 'Agent', and 'Response' blocks.](https://ss.rapidrecap.app/screens/jyWbD_Tp-EA/00-00-07.png)
![Screenshot at 00:28: A chat interface is displayed, showing a conversational AI assistant with a prompt input field.](https://ss.rapidrecap.app/screens/jyWbD_Tp-EA/00-00-28.png)
![Screenshot at 00:57: The video shows the process of cloning the SIM repository from GitHub using the command line interface.](https://ss.rapidrecap.app/screens/jyWbD_Tp-EA/00-00-57.png)
![Screenshot at 01:18: The file explorer within a code editor displays the directory structure of the SIM project, highlighting key configuration files like '.env.example'.](https://ss.rapidrecap.app/screens/jyWbD_Tp-EA/00-01-18.png)
![Screenshot at 01:39: The '.env.example' file is shown, detailing the configuration variables required for setting up the application, including database URLs and API keys.](https://ss.rapidrecap.app/screens/jyWbD_Tp-EA/00-01-39.png)
![Screenshot at 01:57: The terminal displays the output of pulling an AI model \('llama3.2'\) using Ollama, indicating successful download and verification.](https://ss.rapidrecap.app/screens/jyWbD_Tp-EA/00-01-57.png)
![Screenshot at 02:09: A Docker command is executed in the terminal to run a PostgreSQL database container, named 'simstudio-db'.](https://ss.rapidrecap.app/screens/jyWbD_Tp-EA/00-02-09.png)
![Screenshot at 02:31: The 'docker ps' command output confirms that the PostgreSQL container is running.](https://ss.rapidrecap.app/screens/jyWbD_Tp-EA/00-02-31.png)
![Screenshot at 02:45: The installation command for the Bun JavaScript runtime is shown in the terminal, demonstrating a one-line installation process from the command line.](https://ss.rapidrecap.app/screens/jyWbD_Tp-EA/00-02-45.png)