From 3bca3814d91444b6c4ef8ae20a4d776807bf4564 Mon Sep 17 00:00:00 2001 From: Michael Winczuk Date: Wed, 8 Apr 2026 18:43:53 -0700 Subject: [PATCH] Add Math Swarm to LLM tools list --- llm-tools.md | 1 + 1 file changed, 1 insertion(+) diff --git a/llm-tools.md b/llm-tools.md index d7d8583..9594a19 100644 --- a/llm-tools.md +++ b/llm-tools.md @@ -161,6 +161,7 @@ - [BriefGPT](https://github.com/e-johnstonn/BriefGPT) document summarization and querying using OpenAI' and locally run LLM's using LlamaCpp or GPT4ALL, and embeddings stored as a FAISS index, built using Langchain. - [anything-llm](https://github.com/Mintplex-Labs/anything-llm) end to end production ready RAG supports multiple vector DBs, remote and local LLMs and supports chat and query mode with Chat Web UI, agents, code execution, web browsing, multi user, citations, multi user, docker - [factool](https://github.com/gair-nlp/factool) factuality Detection in Generative AI +- [Math Swarm](https://github.com/michaelwinczuk/math-swarm) zero-hallucination computation engine replacing LLM token prediction with SymPy symbolic math. 1,079 tests, 12 categories, 100% accuracy. Benchmarked: 3B (55%) vs 7B (77%) vs 32B (93%) vs Math Swarm (100% at 1.9ms, $0). 15 healthcare clinical formulas with guideline-based decision support - [opencopilot](https://github.com/opencopilotdev/opencopilot) LLM agnostic, open source Microsoft Copilot alternative to easily built copilot functionality with RAG, Knowledgebase, Conversional History, Eval, UX into your product - [DocsGPT](https://github.com/arc53/DocsGPT) chat with your project documentation using RAG, supports OpenAI and local LLMs, and also provides a RAG-fine-tuned docsgpt-14b [model](https://huggingface.co/Arc53/docsgpt-14b) - [Swiss Army Llama](https://github.com/Dicklesworthstone/swiss_army_llama) FastAPI service for semantic text search using precomputed embeddings and advanced similarity measures, with built-in support for various file types through textract