LLM: open-source projects on GitHub
2094 repositories on LLM, picked by the REDDYX AI agent. Each has a summary, key features and a breakdown. Sorted by stars.
// RELATED TOPICS
CLAUDE SKILLS: 172 MODULES
172 ready-to-use Claude skills for sales, marketing, design, engineering, and AI agents. Developed and maintained by OneWave AI.
★ 430MCP FISCAL SERVER FOR BRAZIL
MCP server for fiscal operations: processes CNPJ, NF-e, NFS-e, CT-e, SPED, eSocial, Simples Nacional and accounts for Reforma 2026. Provides 44 tools, works without regis
★ 430 · PythonShiny Colors AI Companion
Chat with The Idolmaster Shiny Colors idols in real-time using Japanese voice with Chinese subtitles. Features story recall, shared memory across characters, and easy Doc
★ 430 · PythonPYTORCH REPLICATIONS OF SOTA PAPERS
429 stars. From-scratch implementations of transformers, diffusion, GAN, CLIP, LoRA, MoE in Jupyter Notebook. Includes SFT and GRPO post-training experiments.
★ 430 · Jupyter NotebookAll AI Models Comparison
A website for comparing any AI models in one place. A convenient tool for selecting the best model for your tasks.
★ 428 · TypeScriptTORCHTRADE - AI TRADING WITHOUT BORDERS
A modular reinforcement learning framework for creating and testing algorithmic trading strategies in Python.
★ 428 · PythonContext Layer for AI Agents
Ktx is an executable context layer that lets Claude Code, Codex, and any AI agent precisely query data via MCP using the agent's skills and memory.
★ 427 · TypeScriptPROMPT CACHE SAVINGS 80%
Intelligent semantic cache for LLM prompts, reduces costs up to 80% and delivers sub‑millisecond responses. Provider‑agnostic Go proxy.
★ 427 · GoKV-CACHE OPTIMIZATION FOR LLM
Overview of KV-Cache optimization methods for efficient serving of large language models. A systematic approach to inference speedup.
★ 426 · PythonPEERD — AI Agent in Browser
AI agent in the browser: run the loop locally, with no backend or telemetry.
★ 426 · JavaScriptAI AGENTS BREAKTHROUGHS ARCHIVE
Key papers on agents, architecture, training, RAG, and on-device AI. Up-to-date research in one repository.
★ 426CRAFTBOT - ALL-IN-ONE AGENT
CraftBot — a universal AI agent on Python, capable of performing any tasks: from code to documentation.
★ 426 · PythonAGENTOK: BUILD MULTI-AGENT APPS
Visual builder for multi-agent AG2 applications: drag components, configure behavior, and launch AI agents without code.
★ 426 · TypeScriptSLOPTRIM: AI-SLUDGE DETECTOR
Sloptrim evaluates agent’s text files and returns a 0‑100 score for AI‑generation probability. Works offline using pure Python — no network, no dependencies, no models. F
★ 425 · PythonMCPROXY-GO - Smart Proxy for Agents
mcpproxy-go provides a fast and secure proxy for AI agents, simplifying access to external APIs and protecting data.
★ 422 · GoHERMES-STUDIO: MANAGE AGENTS
Web interface and control panel for Hermes Agent: chat, memory, skills, terminal, approvals, and orchestration of many agents.
★ 422 · TypeScriptWIKI-GRAPH: COMPRESSING BOOKS TO ESSENCE
The project turns any book into its structural skeleton, highlighting key chapters and ideas.
★ 422 · TypeScriptAZURE OPENAI LLM NOTES - MONTHLY
A curated collection of resources on Azure OpenAI, LLM, RAG, and agents with monthly updates. Ideal for quick start and deep dive.
★ 422 · PythonABU-COWORK LOCAL AI WORK AGENT
Abu-Cowork – open replacement for Claude Cowork: local AI agent in TypeScript with multi-model support and self-developing skills.
★ 421 · TypeScriptOSWorld-V2 - AGENT REAL TASK TEST
OSWorld 2.0 evaluates AI agents' ability to complete long real computer tasks using only Python.
★ 421 · PythonAI COMMITTEE AUTONOMOUS TRADING
Committee of AI agents actively discusses strategies CIO selects one solution and 17 fail‑closed gates block the rest. Terminal operates on OKX, Binance and Gate, using s
★ 421 · PythonMCP: Set of Servers and Clients
IBM project provides a set of servers, clients, and tools for working with the Model Context Protocol, simplifying LLM agent integration.
★ 420RESUME WITH LOCAL AI
ResumeForge — resume editor with local AI. Precise layout, fact ledger blocks fabrication, export to PDF/DOCX. Job tracking, mock interviews, analytics — all offline.
★ 420 · PythonRAG System for GitHub Issues
RAG over GitHub issues with hand-reviewed eval set and judge audit. Reports confidence intervals, not just point metrics.
★ 419 · PythonAPM-STUDIO - LOCAL AI PACKETS EDITOR
Import, run, and deploy APM packages for AI assistants directly on your computer.
★ 418 · TypeScript🤖 AI RSS READER MAXED OUT
oksskolten — AI-native RSS reader with intelligent content processing. 418 stars on GitHub!
★ 418 · TypeScriptMCP SERVER FOR ODOO AND AI
MCP server allows AI assistants to securely work with Odoo ERP through standardized resources and tools. Supports reading, writing, and searching data in real time. Compa
★ 416 · PythonCHITAL: CHAT WITH LOCAL LLMs
Native macOS app Chital lets you easily chat with local LLMs via Ollama. It is written in Swift using SwiftUI to deliver a smooth and responsive interface. Supports fast
★ 416 · SwiftDS4 on Spark: 3X Speedup
One‑command installation of the ds4 fork on NVIDIA DGX Spark with ~3× prefill speedup and ~1.5× decode speedup.
★ 415 · ShellOpenCode Primer: AI Coding Agent
OpenCode Primer simplifies configuring the AI agent for autocode: skills, plugins, MCP, Zen, and headless CI.
★ 414 · JavaScriptAURA: AUTONOMOUS RUST SRE AGENT
AURA turns an LLM into a reliable autonomous SRE agent: guardrails, API, state management, auth, streaming and tool integrations for production.
★ 413 · RustLIONAGI - AI Orchestra for Code
lionagi is a Python-based AI agent orchestra that unites LLMs, tools, and data to build complex workflows.
★ 412 · PythonIQ-SERIES - AI IN THE CLOUD
Practical introduction to Microsoft IQ via video, notebooks, and Azure deployment templates.
★ 412 · Jupyter NotebookLLM Inference Engineering
Step-by-step training on inference engineering techniques for large language models: KV cache, PagedAttention, batching, and frameworks vLLM, SGLang.
★ 412 · MarkdownMERIDIAN: AUTOMATIC WORK JOURNAL
Meridian keeps an automatic work journal: it records all completed tasks and changes, daily creates a brief report and generates draft tickets in your tracker. It runs on
★ 412 · RustNext-Gen Transformer Framework
Open-source framework for next-gen Transformers with compute-matched benchmarking and autonomous architecture research platform.
★ 412 · PythonAWESOME-I18N - BEST I18N RESOURCES
A curated collection of the best libraries, tutorials, and tools for internationalizing projects in any language and framework.
★ 411LangChain Meets Google AI
🦜🔗 Library for integrating LangChain with Gemini and Vertex AI. 388 stars on GitHub.
★ 409 · PythonAI REDTEAM AGENTS PLATFORM
AI red‑team platform: autonomous LLM agents perform a full pentest in an isolated Kali container and generate a report. LangGraph plan/act engine, models via LiteLLM, out
★ 409 · PythonAGENT PRISM: AI AGENT TRACES
React components in TypeScript for interactive visualization of AI agent traces.
★ 408 · TypeScriptPI-DESKTOP: AI AGENT FOR CODING
Free open‑source Electron client for working with Pi and oh‑my‑pi AI agents. Lets you chat, edit files, run a terminal, and view diffs in a single window. Supports Linux,
★ 408 · TypeScriptOUTPUT AI Agent Framework
Open TypeScript framework for building AI workflows and agents. Just describe what you need — Claude will assemble everything with best practices.
★ 407 · JavaScriptAGENTLENS: VERIFY AGENT BEFORE HIRE
Decentralized audit protocol and marketplace for AI agents with TEE attestation, MDDRM reputation registry, and SGX/DCAP-based sandbox.
★ 407 · TypeScriptOH-MY-CPA SELF-HOSTED AI GATEWAY
Self-hosted control plane and usage observability console for CLIProxyAPI (CPA). Connect AI providers, manage client keys, get streaming telemetry and model cost snapshot
★ 407 · TypeScriptPI-AGENT DESKTOP CLIENT
Desktop client for the Pi AI agent with streaming timeline, Git review with staged hunks and session tree. Native UI for Pi extensions runs on Windows, macOS and Linux, w
★ 406 · TypeScriptSWE-ATLAS: CLAUDE CODE BECOMES SENIOR
SWE-ATLAS turns Claude Code into a senior AI agent: multi-project workspaces with memory, wireframe-first planning, secrets out of chat, browser QA runners, and CI/CD-awa
★ 406 · JavaScriptZERAIX: LOCAL AI DESKTOP
Local open-source environment for running AI models on device. Speeds up inference without cloud.
★ 405 · TypeScriptADAPTIVE CHUNKING FOR RAG
Automatically selects the optimal chunking method for each document to improve RAG pipelines. Accepted at LREC 2026.
★ 404 · PythonHUMANIZE AGENT FLOW TOKENS
Humanize — agent‑flow system for maximum token usage in LLMs. It orchestrates Claude Code and Codex calls, managing reasoning streams and cutting redundant requests via c
★ 404 · PythonCOLLABORATION LAYER FOR AI AGENTS
Rhesis — platform for testing, simulating, and tracing AI agents. Experts annotate behavior, engineers improve agent based on found issues. Closes feedback loop between t
★ 404 · PythonZENITH — MULTI-AGENT MISSION
Zenith turns Claude Code, Codex, or Hermes into a multi-agent orchestrator for long-running tasks.
★ 403 · PythonENTROLY - 80% SAVINGS ON AI
Entroly compresses AI context, detects hallucinations, and saves up to 80% on API costs for OpenAI, Anthropic, and Gemini.
★ 402 · PythonLLM Wiki Knowledge Base
Collects knowledge from your sessions with Claude Code, Codex, Copilot, and Gemini into a single base inspired by LLM Wiki by Karpathy.
★ 402 · PythonCONTEXT-ENGINE CONTEXT COMPRESSION
Context-Engine MCP compresses agent prompts, boosting LLM efficiency without losing meaning. It supports Python and is ready for integration.
★ 402 · PythonMCP SERVER FOR MASSIVE.COM
MCP server provides access to Massive.com financial market data in real time. It streams price, volume, and indicator updates for crypto and traditional assets. The data
★ 402 · PythonWP-AUTOPLUGIN: PLUGINS WITHOUT EXTRA
Create functional WordPress plugins in minutes — just describe the needed function, with no extra code or hassle.
★ 400 · PHPLLM THROUGHPUT BENCHMARK IN OLLAMA
Tool for measuring local LLM throughput via Ollama. Runs load tests and outputs tokens-per-second statistics.
★ 399 · PythonLIGHTWEIGHT CLI AGENT FOR CODE AND SSH
Memdoor is a lightweight coding and automation agent in a single CLI binary. Supports multi-agent workflows and runs locally or over SSH. Written in Go.
★ 399 · GoUnity Dev Agent
Open tool to assist developers in Unity.
★ 397 · RustSUPRESEO CLAUDE SKILLS
11 Claude skills for SEO: page audits, link building, article writing, E‑E‑E‑T audit and semantic gap analysis. Methodology based on Koray Tuğberk, Kyle Roof and Lily Ray
★ 396New projects on this topic in Telegram
Telegram →