Vosk Lightweight open-source offline ASR toolkit supporting many languages and on-device use. 01020 Speech-to-Text# ASR# multilingual# offline
Danswer Open-source enterprise search & chat over docs, respecting permissions. 0980 Q&A over Files (RAG)# Danswer# enterprise search# open source
OpenDevin (Open Source) Community-driven project to build an open autonomous AI software engineer. 0830 Code Assistants# agent# automation# open source
Cline (Open Source) Autonomous coding agent for VS Code—plan/act modes, terminal execution, MCP, bring-your-own model. 0680 Code Assistants# agent# BYO model# MCP
SuperAGI Dev-first open-source autonomous agent framework for building and running multi-step agents. 0540 Multistep Agents# agents# autonomous# framework
OpenHands (formerly OpenDevin) Open-source platform for autonomous software development agents that execute multi-step work. 0540 Multistep Agents# coding agent# multi-step# open source
doccano (OSS) Open-source text annotation for classification, NER, and seq2seq tasks. 0470 Datasets & Labeling# doccano# NER# open source
Continue (Open Source) Open-source IDE agent framework; build custom AI code agents for VS Code/JetBrains. 0470 Code Assistants# agents# CLI# Continue
Qwen2.5-Coder (Open Models) Code-specialized model family (0.5B–32B) with competitive generation and reasoning. 0460 Code Assistants# code LLM# generation# open source
LLaVA-OneVision 1.5 Fully open multimodal (images/video+text) models & training stack; strong results and reproducible recipes. 0430 Open-source Models# LLaVA# local# multimodal
OpenAI Whisper Open-source multilingual ASR model known for robustness on diverse audio and accents. 0430 Speech-to-Text# multilingual# offline# open source
NVIDIA NeMo ASR Open toolkit with SOTA ASR models (Conformer, CTC/Transducer) for training and deployment. 0420 Speech-to-Text# ASR models# Conformer# NeMo
Label Studio (OSS) Open-source, highly configurable labeling for text, images, audio, video, and chat. 0400 Datasets & Labeling# active learning# labeling# multimodal
DeepSeek-R1 (and distilled checkpoints) Open (MIT-licensed) reasoning model with distilled 1.5B–70B local checkpoints; known for chain-of-thought quality. 0390 Open-source Models# DeepSeek-R1# distilled# local
RWKV-7 (family) Attention-free RNN-style open models with transformer-level performance and constant-memory inference. 0380 Open-source Models# linear time# local# open source