Today's LLM & Language Models: Fastest-Growing Projects — April 19, 2026
Today's the LLM & Language Models space, we're seeing a surge in tools that focus on making large language models more accessible and user-friendly. Many of these projects aim to simplify the process of interacting with LLMs, whether it's by creating more natural-sounding responses or providing easy-to-use APIs. Meanwhile, others are building on top of existing LLM frameworks to create new applications.
arman-bd/guppylm is a ~9M parameter LLM that talks like a small fish, with a Growth Score of 91.07 and 2,937 stars. Its unique personality has captured the attention of developers, who are contributing to its growth by submitting commits (24 in the past 30 days) to improve its conversational abilities.
hexiecs/talk-normal is another project that's gaining traction, with a Growth Score of 75.82 and 1,239 stars. This system prompt removes AI slop from LLM responses, making them sound more natural and human-like. Its high commit activity (56 in the past 30 days) suggests that developers are actively working to refine its capabilities.
sdyckjq-lab/llm-wiki-skill is a personal knowledge base builder based on Karpathy's llm-wiki method, with a Growth Score of 59.00 and 857 stars. Its support for multiple platforms has made it attractive to users who want to create their own LLM-powered wikis. The project's high commit activity (67 in the past 30 days) indicates that developers are continuously improving its features.
mnfst/awesome-free-llm-apis is a curated list of permanent free LLM APIs, boasting a Growth Score of 55.67 and 2,386 stars. Its popularity stems from the fact that it provides easy access to API keys for various LLMs, making it a go-to resource for developers who want to experiment with different models.
amitshekhariitbhu/llm-internals is an educational project that teaches users about LLM internals step-by-step, with a Growth Score of 52.50 and 549 stars. Its low commit activity (9 in the past 30 days) suggests that the project is relatively stable, but its star count indicates that it's still widely popular among those interested in learning about LLMs.
kessler/gemma-gem runs Google's Gemma 4 model entirely on-device via WebGPU, with a Growth Score of 41.39 and 769 stars. Its unique approach to running LLMs locally has attracted attention from developers who want to explore alternative deployment methods. The project's moderate commit activity (24 in the past 30 days) suggests that it's still under active development.
Pratiyush/llm-wiki is another implementation of Karpathy's LLM Wiki pattern, with a Growth Score of 34.77 and 123 stars. Its high commit activity (100 in the past 30 days) indicates that the developer is actively working to improve its features. askalf/dario is a local LLM router that provides a single endpoint for multiple providers, boasting a Growth Score of 34.14 and 73 stars. Its high commit activity (100 in the past 30 days) suggests that it's still under heavy development.
lucasastorian/llmwiki is an open-source implementation of Karpathy's LLM Wiki, with a Growth Score of 33.13 and 538 stars. Its moderate commit activity (38 in the past 30 days) indicates that developers are contributing to its growth. atomicmemory/llm-wiki-compiler is a knowledge compiler that creates interlinked wikis from raw sources, boasting a Growth Score of 32.29 and 601 stars. Its moderate commit activity (21 in the past 30 days) suggests that it's still under development.
arman-bd/guppylm is a ~9M parameter LLM that talks like a small fish, with a Growth Score of 91.07 and 2,937 stars. Its unique personality has captured the attention of developers, who are contributing to its growth by submitting commits (24 in the past 30 days) to improve its conversational abilities.
hexiecs/talk-normal is another project that's gaining traction, with a Growth Score of 75.82 and 1,239 stars. This system prompt removes AI slop from LLM responses, making them sound more natural and human-like. Its high commit activity (56 in the past 30 days) suggests that developers are actively working to refine its capabilities.
sdyckjq-lab/llm-wiki-skill is a personal knowledge base builder based on Karpathy's llm-wiki method, with a Growth Score of 59.00 and 857 stars. Its support for multiple platforms has made it attractive to users who want to create their own LLM-powered wikis. The project's high commit activity (67 in the past 30 days) indicates that developers are continuously improving its features.
mnfst/awesome-free-llm-apis is a curated list of permanent free LLM APIs, boasting a Growth Score of 55.67 and 2,386 stars. Its popularity stems from the fact that it provides easy access to API keys for various LLMs, making it a go-to resource for developers who want to experiment with different models.
amitshekhariitbhu/llm-internals is an educational project that teaches users about LLM internals step-by-step, with a Growth Score of 52.50 and 549 stars. Its low commit activity (9 in the past 30 days) suggests that the project is relatively stable, but its star count indicates that it's still widely popular among those interested in learning about LLMs.
kessler/gemma-gem runs Google's Gemma 4 model entirely on-device via WebGPU, with a Growth Score of 41.39 and 769 stars. Its unique approach to running LLMs locally has attracted attention from developers who want to explore alternative deployment methods. The project's moderate commit activity (24 in the past 30 days) suggests that it's still under active development.
Pratiyush/llm-wiki is another implementation of Karpathy's LLM Wiki pattern, with a Growth Score of 34.77 and 123 stars. Its high commit activity (100 in the past 30 days) indicates that the developer is actively working to improve its features. askalf/dario is a local LLM router that provides a single endpoint for multiple providers, boasting a Growth Score of 34.14 and 73 stars. Its high commit activity (100 in the past 30 days) suggests that it's still under heavy development.
lucasastorian/llmwiki is an open-source implementation of Karpathy's LLM Wiki, with a Growth Score of 33.13 and 538 stars. Its moderate commit activity (38 in the past 30 days) indicates that developers are contributing to its growth. atomicmemory/llm-wiki-compiler is a knowledge compiler that creates interlinked wikis from raw sources, boasting a Growth Score of 32.29 and 601 stars. Its moderate commit activity (21 in the past 30 days) suggests that it's still under development.