OpenViSTAI · Agents · Blockchain · Crypto
Submit a project
Menu
Agent Frameworks

AI Agents From Scratch

AI Agents From Scratch is an MIT-licensed JavaScript repository that teaches developers to build AI agents locally with local LLMs and node-llama-cpp, progressing from basic inference to tools, memory, ReAct loops and planning patterns.

Explore the practical guide ↓

Independent OpenViST project profile, based on linked sources.

AI Agents From Scratch — OpenViST project profile in Agent Frameworks.
OpenViST project illustration. See the linked sources for product details.
Understand the project

Practical overview

When it fits

Use it when the goal is to understand or prototype agent mechanisms locally before working with higher-level frameworks.

Who it is for

  • Developers learning AI agent architecture from first principles
  • JavaScript builders exploring local LLM inference
  • Teams comparing agent patterns before adopting a production framework

Documented uses

  • Study how local LLM calls become tool-using agents
  • Build educational prototypes for function calling and JSON Schema tools
  • Explore persistent memory and retrieval strategies
  • Experiment with ReAct reasoning-and-acting loops
  • Compare agent architectures through progressively structured examples
Before you choose

Requirements & limits

The repository is public and identifies an MIT license. Its documented setup uses Node.js 18+, downloaded local models and command-line scripts.

  • Local model files are required for the main learning path.
  • The repository uses JavaScript and node-llama-cpp rather than an agent framework.
  • The examples progress from inference and prompting to tools, memory, ReAct and planning.
  • The documentation also includes an optional example for calling hosted models, including GPT-4.
  • The local execution approach is documented, but operating-system-specific setup details are not provided.

Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.

Source-based practical guide

Set up the local agent-learning examples

Install the repository and prepare its documented prerequisites for running the example scripts.

Before you start

  • Node.js 18 or newer
  • At least 8GB of RAM; 16GB recommended
  • Downloaded models placed in the ./models/ folder
  • A Linux, macOS or Windows environment with CLI access

Steps

  1. Download the models and place them in the ./models/ folder, following DOWNLOAD.md for model details.

  2. Install the JavaScript dependencies with npm install.

  3. Run the introductory script with node intro/intro.js.

  4. Run the simple agent and ReAct examples with node simple-agent/simple-agent.js and node react-agent/react-agent.js.

Expected outcome

The documented introductory, simple-agent and ReAct example scripts are set up for study and local execution.

Guide limitations

  • The available documentation does not provide operating-system-specific setup details.
  • The repository presents these as learning examples rather than a production deployment procedure.
Documentation 1 ↗
Evidence and transparency

Sources & review

Official repository ↗
Source record details
Source availability
Active
Sources checked
2 October 2026
Original website title
pguso/ai-agents-from-scratch · Official homepage title · retained in its source language
How we review sourcesSuggest a correction