Today's LLM & Language Models: Fastest-Growing Projects — April 19, 2026
Today's the LLM & Language Models space, we're seeing a surge in interest around tools that facilitate more human-like interactions with large language models. Developers are creating innovative solutions to make these powerful AI systems more accessible and user-friendly, from fine-tuning their output to creating personal knowledge bases. The trend is clear: people want to harness the potential of LLMs without sacrificing readability or usability.
arman-bd/guppylm has taken the top spot with a growth score of 90.83 and over 2,933 stars. This ~9M parameter LLM talks like a small fish, making it an intriguing example of how language models can be designed to mimic specific styles or personas. Its rapid growth suggests that developers are eager to experiment with unique voice options in their own projects.
hexiecs/talk-normal is another standout, boasting a growth score of 74.77 and over 1,222 stars. This system prompt removes AI "slop" from LLM output, allowing users to interact with these models in a more natural way. Its popularity indicates that developers are looking for ways to refine their language models' responses to better suit human communication.
sdyckjq-lab/llm-wiki-skill has garnered significant attention with its growth score of 58.36 and over 842 stars. This personal knowledge base construction skill, inspired by Karpathy's llm-wiki method, supports multiple platforms and allows users to build their own LLM-powered wikis. Its growth suggests that developers are interested in harnessing the power of LLMs for knowledge management and organization.
mnfst/awesome-free-llm-apis has attracted over 2,227 stars and boasts a growth score of 52.36. This curated list of permanent free LLM APIs (complete with API keys) provides developers with a valuable resource for integrating language models into their projects without incurring costs. Its popularity highlights the demand for accessible and affordable LLM solutions.
amitshekhariitbhu/llm-internals, with its growth score of 51.50 and over 538 stars, offers a step-by-step guide to learning LLM internals, covering topics from tokenization to attention to inference optimization. This repository's growth indicates that developers are seeking to deepen their understanding of how language models work under the hood.
kessler/gemma-gem has gained traction with its growth score of 40.71 and over 759 stars. This on-device implementation of Google's Gemma 4 model uses WebGPU, eliminating the need for API keys or cloud services. Its popularity suggests that developers are interested in exploring alternative deployment options for LLMs.
Pratiyush/llm-wiki has seen significant growth with its score of 34.59 and over 119 stars. This repository implements Karpathy's LLM Wiki pattern, allowing users to create a knowledge base from their Claude Code, Codex CLI, Copilot, Cursor & Gemini sessions. Its growth indicates that developers are looking for ways to harness the power of LLMs for personal knowledge management.
askalf/dario has attracted attention with its growth score of 34.14 and over 73 stars. This local LLM router provides a single endpoint for multiple providers, allowing users to access various language models without changing their tools. Its popularity highlights the demand for flexible and adaptable LLM solutions.
lucasastorian/llmwiki boasts a growth score of 32.60 and over 528 stars. This open-source implementation of Karpathy's LLM Wiki enables users to upload documents, connect their Claude account via MCP, and generate wiki content. Its growth suggests that developers are interested in creating personalized knowledge bases using LLMs.
atomicmemory/llm-wiki-compiler has gained traction with its growth score of 31.57 and over 587 stars. This knowledge compiler transforms raw sources into interlinked wikis, inspired by Karpathy's LLM Wiki pattern. Its popularity indicates that developers are seeking tools to help them organize and structure their knowledge using LLMs.
arman-bd/guppylm has taken the top spot with a growth score of 90.83 and over 2,933 stars. This ~9M parameter LLM talks like a small fish, making it an intriguing example of how language models can be designed to mimic specific styles or personas. Its rapid growth suggests that developers are eager to experiment with unique voice options in their own projects.
hexiecs/talk-normal is another standout, boasting a growth score of 74.77 and over 1,222 stars. This system prompt removes AI "slop" from LLM output, allowing users to interact with these models in a more natural way. Its popularity indicates that developers are looking for ways to refine their language models' responses to better suit human communication.
sdyckjq-lab/llm-wiki-skill has garnered significant attention with its growth score of 58.36 and over 842 stars. This personal knowledge base construction skill, inspired by Karpathy's llm-wiki method, supports multiple platforms and allows users to build their own LLM-powered wikis. Its growth suggests that developers are interested in harnessing the power of LLMs for knowledge management and organization.
mnfst/awesome-free-llm-apis has attracted over 2,227 stars and boasts a growth score of 52.36. This curated list of permanent free LLM APIs (complete with API keys) provides developers with a valuable resource for integrating language models into their projects without incurring costs. Its popularity highlights the demand for accessible and affordable LLM solutions.
amitshekhariitbhu/llm-internals, with its growth score of 51.50 and over 538 stars, offers a step-by-step guide to learning LLM internals, covering topics from tokenization to attention to inference optimization. This repository's growth indicates that developers are seeking to deepen their understanding of how language models work under the hood.
kessler/gemma-gem has gained traction with its growth score of 40.71 and over 759 stars. This on-device implementation of Google's Gemma 4 model uses WebGPU, eliminating the need for API keys or cloud services. Its popularity suggests that developers are interested in exploring alternative deployment options for LLMs.
Pratiyush/llm-wiki has seen significant growth with its score of 34.59 and over 119 stars. This repository implements Karpathy's LLM Wiki pattern, allowing users to create a knowledge base from their Claude Code, Codex CLI, Copilot, Cursor & Gemini sessions. Its growth indicates that developers are looking for ways to harness the power of LLMs for personal knowledge management.
askalf/dario has attracted attention with its growth score of 34.14 and over 73 stars. This local LLM router provides a single endpoint for multiple providers, allowing users to access various language models without changing their tools. Its popularity highlights the demand for flexible and adaptable LLM solutions.
lucasastorian/llmwiki boasts a growth score of 32.60 and over 528 stars. This open-source implementation of Karpathy's LLM Wiki enables users to upload documents, connect their Claude account via MCP, and generate wiki content. Its growth suggests that developers are interested in creating personalized knowledge bases using LLMs.
atomicmemory/llm-wiki-compiler has gained traction with its growth score of 31.57 and over 587 stars. This knowledge compiler transforms raw sources into interlinked wikis, inspired by Karpathy's LLM Wiki pattern. Its popularity indicates that developers are seeking tools to help them organize and structure their knowledge using LLMs.