scenario-labs / blender-plugin
Scenario for Blender: AI images, video, 3D and PBR materials in the viewport, plus a local MCP server for agents. Blender 5.0+ extension, GPL-3.0-or-later.
scenario-labs / blender-plugin
fishfinger44 / MatteDesk
Interactive matte / roto tool for VFX – repairs damaged plate frames, generates alpha mattes for multiple objects (SAM3, CoTracker3, MatAnyone2, VideoMaMa, ViTMatte, LTX 2.5 Alpha Gen via ComfyUI), exports to Nuke and After Effects. REST API + CLI for AI agents. GPL-3.0.
rollroyces / py-idp
AI-enabled Intelligent Document Processing for Python: 12+ LLM backends, Pydantic schemas, self-hosted OCR (Nanonets-OCR2-3B), auto-schema discovery, HITL review, prod FastAPI server. AGPL-3.0 + commercial.
bayshier / voicestudio-zh
VoiceStudio 中文版：开源、完全本地的 ElevenLabs 替代品——语音克隆、语音设计、视频配音、听写、转写与有声书，646 种语言（译自 debpalash/VoiceStudio，AGPL-3.0）
albumentations-team / AlbumentationsX
Image augmentation for computer vision. AGPL-3.0-only or commercial licensing.
nullmoth / 1401-probe
Free, offline hardware scanner on a USB stick. Scans a PC's hardware and saves one report; never touches your drives or the network.
orbi-build / orbi
Open-source (AGPL-3.0), self-hostable lights-out software factory for the agent era: label a GitHub Issue, get an independently reviewed, merged PR and a tagged release.
jerrypps3-glitch / chanlun-engine
缠论引擎 V2 — 标准108课算法管线 (分型/笔/段/中枢/BSP), 零第三方依赖, AGPL-3.0
databluedev / searchmirror
Open-source, self-hosted rank tracker and AI-visibility monitor built on the DataBlue SERP API. Bring your own key. AGPL-3.0.
EXwhitewood / wanxiangpu-loreweft
万象谱 Loreweft｜面向长篇小说工程的 AI 原生创作工作台｜AGPL-3.0 开源
MarkYoungIV / abacuscopilot
Interactive command-line pre/post-processing copilot for the ABACUS DFT software (GPL-3.0)
Technicalflight / image-prompt-reverse
把一张图片逆推成生成提示词（Image to Prompt）：中英双语重建提示词 + 风格标签 + 结构化 JSON prompt，零依赖校验脚本 | Forensic image-to-prompt reconstruction. AGPL-3.0 · Commercial use requires license.
rufeng0411 / Nova-TradingAgent
Self-hosted A-share multi-agent research desk: 15-agent debate, optional Tushare Level-2 order queue, K-line workstation with AI insight, and a signal translation layer so models read conclusions — not raw dumps. AGPL-3.0.