GitHub - rasbt/LLMs-from-scratch: Implement a ChatGPT-like LLM in PyTorch from scratch, step by step vs GitHub - f/prompts.chat: f.k.a. Awesome ChatGPT Prompts. Share, discover, and collect prompts from the community. Free and open source — self-host for your organization with complete privacy.

## Overview

**LLMs-from-scratch** is an educational GitHub repository that teaches you to implement a ChatGPT-like large language model in PyTorch from scratch. It covers tokenization, attention, pretraining, finetuning, and instruction tuning.

**prompts.chat** is a free, open-source community platform for sharing, discovering, and collecting prompts for LLMs like ChatGPT, Claude, and Gemini. It is self-hostable and focuses on prompt engineering rather than model building.

## Feature Comparison

| Feature | LLMs-from-scratch | prompts.chat |

|---------|-------------------|--------------|

| **Primary Focus** | Building LLMs from scratch | Sharing and discovering prompts |

| **Skill Level** | Intermediate to advanced (requires PyTorch) | Beginner-friendly |

| **Open Source** | Yes (MIT license) | Yes (MIT license) |

| **Self-Hostable** | No (code is local) | Yes |

| **Community Contributions** | Limited to issues/PRs | Active prompt sharing |

| **Multi-Model Support** | GPT-like only | ChatGPT, Claude, Gemini, etc. |

| **Learning Resource** | Comprehensive tutorial | Not educational per se |

## Pros and Cons

### LLMs-from-scratch

**Pros:**

- Deep understanding of LLM internals

- Step-by-step, well-documented code

- Covers full pipeline from tokenization to instruction tuning

- Free and open-source

**Cons:**

- Requires PyTorch knowledge

- Training larger models needs GPU resources

- Not a production-ready framework

- Limited to small-scale models by default

### prompts.chat

**Pros:**

- Completely free and open source

- Self-hostable for privacy

- Large community-driven prompt library

- Multi-model support

- No registration required for browsing

**Cons:**

- Requires GitHub account to contribute prompts

- Self-hosting needs server resources

- Prompt quality may vary

- Limited advanced filtering options

- No official mobile app

## Use Cases

- **If you want to learn how LLMs work internally**, choose **LLMs-from-scratch**. It's ideal for students, researchers, and developers who want to build and modify a GPT-like model from the ground up.

- **If you need ready-to-use prompts for content creation, coding, or education**, choose **prompts.chat**. It's perfect for writers, marketers, and educators who want to leverage LLMs without technical depth.

- **For organizations needing a private prompt library**, prompts.chat's self-hosting capability is a clear advantage.

## Pricing

Both tools are **free and open-source**. LLMs-from-scratch may incur cloud costs if you train on GPU instances. prompts.chat requires server resources for self-hosting but no software fees.

## Conclusion

LLMs-from-scratch and prompts.chat serve different purposes. If your goal is to understand and build LLMs, go with LLMs-from-scratch. If you want to collect and share prompts for everyday use, prompts.chat is the better choice. Both are valuable additions to any AI toolkit.