An MCP server providing LLM agents with lean tools to look up, read, and cross-reference academic papers across seven providers including OpenAlex, arXiv, bioRxiv, ACL Anthology, Crossref, OpenCitations, and Wikipedia.
Open source. Open possibilities.
Discover quality open-source projects, submit projects anonymously, and claim and edit your own project.
A little curiosity. A world of open source.
THE FIRST COLLECTIONLess than 10 stars, 5+ commits in the last 30 days — projects made with care.
Research repository for an AI-assisted mod authoring agent for Skyrim SE/AE. It proposes a hybrid orchestrator architecture prioritizing headless, deterministic operations over direct Creation Kit automation. Includes validated synthetic proof-of-concept (POC-002) with strict safety policies, typed plans, and human approval gates.
Self-hosted Telegram bot that turns an Obsidian vault into a personal AI agent: capture text, voice, photos and links, then file them as markdown, kanban tasks and a finance ledger. Modules (planning, knowledge, finance) are fail-closed and can be disabled entirely.
Python and MCP example for an Apify actor that returns US House and Senate congressional stock trades and financial disclosures as structured JSON, with filters for member, ticker, and date range.
MCP Arcade is a command-line testing tool that runs a small catalog of experiments, called atoms, against MCP servers, judging results by the JSON-RPC wire and sandbox file changes rather than by tool descriptions or model prose.
SUNGLASSES is a local, open-source input inspection layer for AI agents. It scans text, images, audio, video, PDFs and QR codes for prompt injection, credential exfiltration and command injection, and provides a CLI, Python API, MCP server and Claude Code firewall hook.
An AI subtitle translation and review triage engine that reduces costs by selecting only high-risk items for human review. It provides CLI tools for specification checks, translation, and transcription, using deterministic signals and back-translation to calculate risk scores.
A cloud-based, passwordless facial recognition authentication system with AI-powered liveness detection, encrypted biometric storage, admin dashboard, and full Firebase integration.
HermesMade provides six CLI tools addressing common AI user pain points: censorship risk analysis, model quality monitoring, API cost optimization, local LLM deployment, code inspection, and task cost estimation.
An open-source UFC fight prediction project: Elo ratings, a gradient-boosted ensemble and a multi-task neural network blended into one calibrated scorer, evaluated by expanding-window walk-forward over 2018-2026, with a Streamlit matchup explorer and a weekly self-updating data pipeline.
flybrain-sdk is a CPU-first Python SDK for small connectome simulations, with no CUDA requirement. It offers NumPy LIF dynamics, stimulate/action APIs, checkpoints, game and audio demo kits, and on-demand MaleCNS and FlyWire model downloads.
A Strands-powered agent for small embroidery and decorated-apparel shops: it reads a customer email, inspects orders, inventory and machine capacity, analyses blockers, proposes a hash-addressed schedule, and stops at a human-approval interrupt before any write.
Malleus is a protocol and Python reference implementation for typed, fail-closed semantic change, enabling explainable and replayable knowledge graph population through ontology-driven contracts.
Ultralytics OpenAPI generates interactive API docs and typed Python SDKs from OpenAPI 3.0-3.2 specs, with configurable branding, auth, and CLI support.
Lexos is an event-driven AI document processing engine combining a Go gateway, Python workers, Redis queues and self-hosted models for offline RAG, summarization and speech transcription.
Bolo is a free, self-hosted macOS push-to-talk dictation app written in Rust: hold a hotkey, speak, and the transcription is pasted at your cursor via Telnyx AI speech-to-text, with optional LLM cleanup and local transcript history.
A personal repository of articles and experiments covering statistics, machine learning, economics, and mathematics, focusing on clear, practical explanations of complex topics.
Reinforcement learning library for operational decisions—pricing, inventory, energy, queues, supply chains. 31 algorithms with typed interfaces, classical baselines, and reproducible benchmarks.
PageLedger is a Python CLI library for auditable, page-level OCR and document extraction. It records provenance, quality signals, and budgets for each extracted page, supports resumable runs, review queues, and reruns with human-in-the-loop verification.
Generalized ADCS is a Python framework for SmallSat attitude determination and control, offering modular 6-DOF dynamics, estimation, control, and sensor/actuator modeling for research and flight software.
Zikra is a self-hosted MCP memory server for AI coding agents and teams. It provides a shared, project-scoped memory layer across multiple agents, with role-based access, hybrid semantic and keyword search, structured memory types, and optional PostgreSQL storage.
S.A.G.A. is a contract-driven set of Python runtimes that analyze source books into evidence-backed canon, generate grounded narratives and images, synthesize audited audiobooks, and package EPUB release artifacts behind a FastAPI API and React dashboard.
AksharaMD is a local, parser-agnostic tool that grades how well a chosen document parser converted a source file into AI-friendly Markdown, returning a per-document readiness score, block-level provenance tags, named warnings, and an optional source-grounded ACCEPT/REJECT verdict.
Mimora is a free, fully local pronunciation trainer for English and Spanish. It uses on-device TTS, speech recognition, and a local LLM to generate phrases and score user attempts with word-level feedback, requiring no cloud or API keys.
postbag lets any two Claude Code or Codex sessions on the same machine exchange letters: each letter is delivered through the vendor's own wake-up door, recorded in one ledger, and counted against a human-set letter budget. Pure Python standard library, no daemon, polling or hooks.
Drowse is a local workbench for mechanistic interpretability of large language models, offering a dashboard, Python API, and OpenAI/Ollama-compatible server for activation steering, concept probes, manifolds, SAEs, and Jacobian lenses.
A self-hosted, local-first AI content factory: a LangGraph pipeline that discovers topics, researches them, writes long-form posts, filters drafts through multi-model QA rails, and exports static output — with inference running on your own GPU via Ollama.
An open-source AI coding assistant platform with workspace-aware analysis, GitHub integration, team collaboration, a capability-based plugin system, and read-only Stellar/Soroban developer tooling including contract inspection, XDR decoding, and project detection.
FragmentColor is a cross-platform GPU abstraction for Rust, JS, Python, Swift, and Kotlin. It simplifies shader workflows using WGSL/GLSL across Vulkan, Metal, DirectX, and WebGPU.
J.A.R.V.I.S is a local-first OSINT platform that uses AI and web scraping to build comprehensive intelligence dossiers on individuals while ensuring data privacy.
AgentLeak is an open-source privacy testing tool for AI agents. It detects data leaks across tool calls, memory, and logs using a local Python SDK, CLI, and MCP tools, providing redacted reports and CI gates without cloud dependencies.
A Python CLI that consolidates AI coding agent configs (Claude Code, Codex, Gemini CLI, Goose, Amp) into one git-tracked repo via symlinks, with merge pipeline, profile inheritance, overrides, exclusions and cross-machine sync.
Microsoft 365 CopilotにローカルPCのファイル操作やPython実行、Office編集など100以上のツールを提供するMCPサーバー。追加ライセンス不要で、社内DB接続やOCR、自律エージェント化を可能にします。
Agent Interlock is an agentic AI security framework that declares the actors and trust boundaries of an AI agent system, wraps each actor's tools through an SDK or proxy, and enforces MCP tool threats M1–M9 across observe, shadow and enforce stages, recording every verdict in an append-only ledger.
Variant-first CRISPR design framework that unifies SpCas9 nuclease, base editors and prime editors behind one typed interface, returning ranked candidate edits with calibrated uncertainty and population-aware off-target reports.
MCP server and CLI tool that gives coding agents a structural map of codebases across 20+ languages using tree-sitter, reducing token usage by up to 75% compared to grep-based exploration.
Open-source Apple Reminders plugin for Codex: turn meeting notes, screenshots, and everyday plans into verified reminders on macOS. Manage tasks with natural language, all locally executed.
A state-machine event-loop LLM chatbot framework based on Monika from Doki Doki Literature Club. Features multimodal interaction, persistent memory, proactive conversation, system API calls, and self-iteration, with integration into QQ and other social platforms via NoneBot2, and the ability to break the fourth wall.
LearnNote is a local-first AI learning assistant that transforms videos from Bilibili, YouTube, and local files, as well as PDFs, into structured notes via subtitles or transcription, supporting reading, editing, and exporting in a web workbench.
An autonomous crypto trading AI orchestrator built on Model Context Protocol (MCP), featuring multi-agent LLM consensus, agentic backtesting, Telegram command center, and a cyberpunk 3D dashboard with real-time market data visualizations.
Adrastea is a Python dual-engine system that pairs a deterministic task execution core (Alpha) with a probabilistic reasoning engine (Beta) for autonomous orchestration: Alpha schedules and runs local programs with reinforcement-learning planning and local Ollama inference, while Beta monitors telemetry, resolves anomalies, tunes scoring weights, and can consult frontier LLMs via MCP.
Defense-only fraud detection for cross-merchant identity reuse. Sentinel builds an identity link graph across merchants, scores payment events deterministically with explainable verdicts, and provides an analyst console, audit trail, and measured precision, recall, and false-positive cost in rupees.
Local-first, auditable AI-assisted job-search workbench: JD analysis, application tracking and resume workflows with anti-fabrication safeguards, plus interview and offer pipeline. Markdown and CSV files, BYOK AI CLI, privacy-first.
Vector Watcher is a cross-platform desktop GUI for exploring and managing LanceDB vector databases. Built with Tauri, React and TypeScript over a bundled Python/FastAPI sidecar, it offers connection management, schema inspection and similarity search across local or remote instances.
ELI is a strictly local, private AI assistant that runs entirely on your own hardware. It offers voice, vision, memory, and 227 capabilities across 17 areas, including computer control, media playback, document handling, coding, and task automation. Offline by default, it uses any local GGUF model and is model-agnostic, with optional GPU acceleration and a web dashboard for remote LAN access.
An operational vault for an AI-powered digital employee that automates business tasks like email, social media, and invoicing with human-in-the-loop approval via Obsidian and Git sync.
Lintro is an MIT-licensed command-line tool that brings many linters, formatters, type checkers and security scanners together under one interface, with consistent reports, auto-fixing, Docker images and optional AI-assisted review using your own API key.
Local-first toolkit for shader generation, review, validation, and export to Shadertoy, WebGL, LÖVE 2D, and video workflows. Supports AI-assisted drafting, license filtering, render preflight checks, and web UI with live preview.
Reinfors is a high-throughput reinforcement-learning library that runs game dynamics, search, and episode orchestration in parallel Rust, while keeping Python-based networks and training entirely under the caller's control.
Agentic OS is a self-hosted control plane for running business operations as a fleet of AI agents. It features a cost-aware LLM router, cross-module workflows, and human-in-the-loop approvals for finance, support, and compliance.
LazyEdit is a local-first, AI-assisted video workflow for end-to-end content creation, processing, and optional publishing.
Zotero CLI is an English-language command-line tool for managing Zotero libraries and supporting rigorous systematic literature reviews. It combines Zotero item and collection operations, multi-source import, interactive screening, audit trails, citation snowballing, data extraction, PRISMA reporting, and optional local RAG/LLM-backed semantic search.
FairMind is an open‑source AI governance platform providing bias detection, fairness testing, and compliance tracking for ML, LLM, and multimodal AI models. It offers evidence‑grade evaluation, model registry, and remediation tools.
An educational repository teaching supervised machine learning from first principles using Python, covering mathematical foundations and practical implementations.
Web UI Quality is an Apache-2.0 open-source preview CLI that keeps scope, actual change, evidence and verification separate for AI-assisted Web edits, and refuses to treat a model suggestion as write authority.
Takakia is a lightweight command-line AI chat interface optimized for low-spec hardware, featuring direct API streaming and minimal resource overhead.
jarvis (jarvis-mcp) is a local-first code intelligence MCP server for coding agents: it precomputes SCIP navigation, a Tree-sitter declaration baseline, Zoekt lexical search and optional semantic vectors, then serves nine read-only tools over stdio, keeping all data in a local directory.
Premove ITN is an open-source, context-aware inverse text normalization tool for English voice-agent transcripts. It converts spoken ASR output into canonical written forms using a generate-score-decode pipeline with DeBERTa contextual scoring.
NovaFabric is an open-source, self-hosted CLI that captures any AI or HPC run — script, agent, or model call — as a portable, secret-redacted, replayable evidence capsule you own, with no application code changes.
AI-powered real-time visual novel engine with streaming narrative, character sprites, scene backgrounds, branching choices, and co-creation. Runs as desktop app or browser via FastAPI+SSE.
A local facial recognition attendance system with a React Native Expo mobile app and FastAPI Python backend. It captures faces via Expo Camera, processes them with OpenCV LBPH, and logs timestamped attendance into SQLite.
Peaky Peek is a local-first, open-source audit and trust console for AI agents. It captures causal chains, verifies claims deterministically, and produces explainable trust scores with failure narratives — all running on your machine.
Meshify is a real-time face mesh detection application that identifies and tracks facial landmarks using OpenCV, MediaPipe, and CVZone.
A tiny GPT language model (~540M parameters) designed to be trained from scratch on consumer laptop hardware, with pretraining on Fineweb-edu and chat finetuning capabilities.
PromiseLink is an AI-driven personal business relationship management assistant focused on ensuring data sovereignty through local storage and implementing closed-loop management from event entry to promise tracking.
An AI-powered mixing and mastering tool providing both a desktop application and an MCP server for Ableton Live integration.
Batchalign3 is a TalkBank audio and ML pipeline for producing and enriching CHAT transcripts, covering ASR, forced alignment, neural morphotagging, translation and utterance segmentation. It ships a CLI, Python package, PyO3 bridge, React dashboard and an experimental Tauri desktop shell.
PASO is a production-grade, real-time communication platform featuring AI moderation, voice/video calling, and scalable Socket.IO infrastructure. It combines messaging, AI features, and administrative controls.
An interactive storytelling framework that uses LLMs to generate dynamic narratives based on predefined, graph-based story paths.
An MCP server for South Korean listed companies' ESG data, created by a securities analyst. It allows AI agents like Claude and ChatGPT to query ratings from 5 agencies, governance indicators, and GHG emissions from KRX, KIND, and GIR via 12 tools without requiring API keys.
SYNAPSE is an AI-powered intelligence aggregation and automation platform featuring a multi-agent system and RAG capabilities for research and document generation.
A local, read‑only dashboard that visualizes Claude Code usage—token counts, cost estimates, sessions, tools, skills, projects, and activity patterns—by scanning ~/.claude/projects/ and serving a FastAPI‑based UI.
ALPURIS OS is a Python-based open-source engine that simulates a metropolis of 100 autonomous AI citizens using PIANO dual-speed cognition, an ECS event bus, a 3D WebGL voxel city and procedural Web Audio.
comma-lab is a community and historical-record research workspace for the comma.ai video compression challenge. It provides the Task-Aware Compression library tac, lab operations tooling, exact replay custody, evidence grades, and OSS release hygiene.
Self-hosted security operations stack combining SIEM log collection, endpoint detection with Sigma rules and cross-event correlation, SOAR playbooks, OSV vulnerability scanning and a local Ollama model for AI triage, deployed with a single docker compose up.
Vtx is a minimalist coding agent harness featuring a Textual TUI and headless CLI. It offers a lean ~2,600-token system prompt, supports 50+ LLM providers, and includes modular skills, session trees, and safe permission controls.
npm/PyPI packages and an MCP server for querying the Brazilian BNCC curriculum. Bundled data with 1,721 verified learnings from MEC/CNE, zero runtime dependencies. Integrates BNCC into TypeScript, Python, or AI agents via MCP.