Lumo
4.0Proton's privacy-first AI chat assistant with zero-access encrypted chats, open-source code, web search and image generation, on web, iOS and Android.
-
-
Orkas
3.7Open-source, local-first desktop app where a team of specialist AI agents takes a chat brief and delivers research, documents, slides, videos and websites.
-
AI Agent Skills
3.6Directory of more than 8,800 community skills, templates and plugins for Claude, Codex and other AI agents, sorted by category from coding and MCP to marketing.
-
Cedille
3.4French-language AI writing assistant built on a custom-trained GPT-J model, made at EPFL Innovation Park in Switzerland.
-
AIimag.es
3.4Free, open-source Windows program from developer Sunija that runs Stable Diffusion locally to generate images from text, no cloud account needed.
-
xTuring
3.5Open-source Python library for fine-tuning, quantizing and running your own LLMs on your data, from Stochastic.
-
Datumbox
3.0Open-source Java machine learning framework and REST API for sentiment analysis, topic classification and text extraction.
-
deepset
4.1Berlin-based maker of the Haystack framework for building, testing, deploying and governing RAG systems and agents on your own infrastructure.
-
Mistral AI
4.5European frontier AI platform with open models, fine-tuning and enterprise deployment.
-
AlbumentationsX
3.6Cited in 2,270+ research papers and used in 40,000+ GitHub repos – the open-source image augmentation library behind winning solutions in 77 AI competitions, NumFOCUS-affiliated.
-
Picovoice
3.6Voice, language, and vision AI that runs entirely on-device, no cloud required – 15 production-ready products trusted by NASA, Adobe, Meta, Netflix, and Ferrari.
-
Evidently AI
3.7LLM-as-a-Judge evaluation with 100+ built-in metrics for hallucinations, toxicity, and PII – open-source, 7,500+ GitHub stars, used by DeepL, Wise, Plaid, and Databricks.
-
Konduit
3.7Run PyTorch, TensorFlow, Keras and ONNX models wherever you need them – cloud, on-premise, edge or mobile – backed by the Eclipse Deeplearning4j ecosystem.
-
Clips AI
3.4Open-source Python library that reads a video's transcript to find the best moments, then uses AI speaker detection to reframe it into vertical clips automatically.
-
PostgresAI
3.9Automated senior-DBA expertise for Postgres – monitoring, health checks, issue detection, zero-downtime upgrades and thin database clones via DBLab.
-
Prismatic
3.7"The AI-enabled embedded iPaaS" – describe an integration in plain language and an AI Copilot builds it, with agent-ready tools exposed via MCP for AI to act directly.
-
LlamaChat
3.2Run LLaMA, Alpaca, and GPT4All locally on your Mac – no cloud, no data leaving your machine, 100% free and fully open-source.
-
LanguageTool
4.0Multilingual grammar, style and tone checking with a paraphrasing tool, free to use and open source at its core.
-
HOL Guard
3.8Local-first runtime firewall for AI coding agents
-
ml5.js
3.7Machine learning for the browser, built for artists and students – pose estimation, image/sound classification, and custom neural nets on TensorFlow.js.
-
Caffe
3.5Berkeley's deep learning framework, still maintained in 2026 – built for expression, speed, and modularity, processing 60M+ images a day on GPU hardware.
-
EasyOCR
4.0Open-source Python OCR reading 80+ languages, including scene text, in three lines of code.
-
Theneo
3.7API documentation written to be consumed by AI agents as well as developers, with an assistant that answers from your own docs.
-
Roboto AI
3.5Roboto Agents automate triage, root-cause analysis, and anomaly detection across your entire robot fleet's data logs – drones, surgical robots, autonomous vehicles.
-
Weaviate
3.7The open-source vector database behind much of today's RAG stack – built-in embeddings, a natural-language Query Agent, tens of thousands of tenants per cluster, SOC 2 and HIPAA certified.
-
Sumo Logic
3.9Cloud log analytics, monitoring and SIEM on one platform, with Dojo AI adding a multi-agent layer over security and operations data.
-
AutoKeras
3.5Open-source AutoML library for Keras – searches model architectures for you across image, text and structured data tasks.
-
UniJump
3.4Cmd+J anywhere on the web to summarize the page, rewrite a sentence, or translate on the spot – free, open source, no signup, stores no personal data.
-
AppTek.ai
3.4Transcribe, translate, understand, and synthesize speech across dozens of languages – ASR, neural machine translation, LLMs, TTS, and NLP in one enterprise language-AI platform.
-
Baserow
3.6Kuma, Baserow's AI Assistant, builds your data structure from a plain-language description – open source, with AI Agents and MCP Server support for ChatGPT, Mistral, and Claude.
-
NewsBlur
3.6An RSS reader with Intelligence Training – you teach it which authors, tags and phrases you care about, and it filters every feed accordingly.
-
Mapbox
4.0Developer platform for maps, navigation and search, with a Location AI line integrating geospatial data into AI applications.
-
NextBrain AI
3.7AutoML runs experimentation 10x faster, document extraction hits 99% clean fields, and a RAG knowledge repository makes your organization's context searchable 24/7.
-
MLflow
3.8The largest open-source AI engineering platform for agents, LLMs, and ML models – 30M+ downloads a month, Fortune 500 production use, backed by the Linux Foundation.
-
Pet Portrait AI
3.1Fine-tuned Stable Diffusion plus human post-processing turns your pet's photo into 10+ styled portraits inspired by pop culture, history, and famous art – ready in 24-48 hours.
-
Prodia
3.9One API across the best open-source image and video models, with sub-second generation and independently benchmarked speed.
-
Open Vibe
3.7Wraps Claude Code or another coding agent with structure for building a software product.
-
Logz.io
4.0Observability built on open source – logs, metrics and traces with OrionIQ, an agentic layer that investigates incidents rather than only alerting on them.
-
Emdash
4.1An open-source desktop environment for running coding agents like Claude Code and Codex in parallel, with scheduled agent work, a built-in preview browser, and shared prompt, skill and MCP management.
-
Metabase
3.8Metabase AI answers natural-language questions on every tier – including the free, self-hosted open-source edition – with inspectable, verifiable queries.
-
Fast.ai
4.6Free deep learning courses and open-source library that teach practical AI without a PhD prerequisite.
-
Kobiton
3.9Mobile app testing on real devices, with a device cloud, AI-augmented testing, no-code validations and DevOps pipeline integration.
-
DeepDetect
3.7An open-source deep learning REST API server, with a Docker-installable platform UI, Jupyter notebooks and GPU support for training and deployment.
-
Eventual AI
3.4Index every camera on your robot fleet, search for rare failure scenarios in plain language, and auto-curate training datasets – built on the open-source Daft data engine.
-
Stockfish
3.8Free, open-source, and the strongest chess engine in the world – NNUE neural network evaluation trained on millions of positions, running efficiently on a CPU.
-
Plotly
3.8Plotly Studio does agentic data analysis – an LLM builds real Plotly charts and Dash apps, backs every step with visible code, and asks when something's ambiguous.
-
Transfers motion from a reference video onto a still character, no rigging.
-
DashVox
3.8Voice-first control of coding agents over SSH, built for hands-free use rather than a screen.
