ai-agents-from-scratch

ai-agents-from-scratch is an agent framework - Demystify AI agents by building them yourself. Local LLMs, no black boxes, real understanding of function calling, memory, and ReAct patterns.

Overview

Learn to build AI agents locally without frameworks. Understand what happens under the hood before using production frameworks.

This repository teaches you to build AI agents from first principles using local LLMs and node-llama-cpp. By working through these examples, you'll understand:

Philosophy: Learn by building. Understand deeply, then use frameworks wisely.

• Node.js 18+ - At least 8GB RAM (16GB recommended) - Download models and place in './models/' folder, details in DOWNLOAD.md

Decision guide: use ToT when you need to search competing paths; use GoT when you need to combine multiple sources into one consistent policy. Compare both in: - ToT concept - GoT concept

What you'll learn: - Precomputing embeddings for short exemplar phrases per tool - Scoring the user message against exemplars (cosine similarity) with a small embedding model - Passing only top-k tools (plus optional always-include tools) into 'session.prompt' - Observing recall failure when k is too small for multi-intent prompts

• ' .js' - The working code example - 'CODE.md' - Step-by-step code explanation - Line-by-line breakdowns - What each part does - How it works - 'CONCEPT.md' - High-level concepts - Why it matters for agents - Architectural patterns - Real-world applications - Simple diagrams

ai-agents-from-scratch is open-source, written primarily in JavaScript, with 4,240 GitHub stars under the MIT license. It was last updated on 2026-05-31.

Key capabilities

From the project's documentation:

• How LLMs work at a fundamental level
• What agents really are (LLM + tools + patterns)
• How different agent architectures function
• Why frameworks make certain design choices
• Explains why each example exists
• Visualizes the learning path from raw LLM calls to full agents

How it fits a local-AI stack

ai-agents-from-scratch runs on your own hardware, so pair it with a model and a GPU sized to your needs. Use the VRAM calculator to pick a model that fits your card, and see what you can run for hardware guidance. Related agent frameworks in the directory:

• Agno
• AutoGPT
• CrewAI
• Genkit

Sources

• Source code & docs: pguso/ai-agents-from-scratch

Stats from GitHub, 2026-06-08.