OPEN SOURCE OBSERVER
从项目热度走向工程判断：追踪开源工具、模型与开发者生态。
01 / CODE
| 排名 | 项目 | 语言 | 今日新增 | 总 Star |
|---|---|---|---|---|
| 01 | ayghri/i-have-adhdA skill to stop your coding agent from burying the answer. ADHD-friendly output. | Python | +3,440 | 41,483 |
| 02 | bilawalsidhu/gods-eye-viewA spy satellite simulator in your browser, except the data is real. Live open source spatial intelligence on a photorealistic 3D globe. | JavaScript | +3,642 | 26,855 |
| 03 | nab138/iloaderUser friendly sideloader | TypeScript | +36 | 2,880 |
| 04 | melgarafael/DeskcommCRMOpen-source AI sales OS — self-hosted CRM with native AI agents + WhatsApp (WAHA). Open alternative to Kommo, Octadesk & Intercom for any business that sells by chat. MCP-ready, multi-tenant, LGPD. | TypeScript | +126 | 1,284 |
| 05 | vastsa/PI-DesktopLocal-first AI coding agent desktop: Electron + Rust host core + pi Agent Harness + user-installable plugins | TypeScript | +545 | 2,748 |
| 06 | armory3d/armorpaintGraphics Creation Tools | C | +354 | 4,696 |
| 07 | alsk1992/CloddsBotOpen Source AI trading agent that operates autonomously across 1000+ markets - Polymarket, Kalshi, Binance, Hyperliquid, Solana DEXs, 5 EVM chains. Scans for edge, executes instantly, manages risk while you sleep. Agent commerce protocol for machine-to-machine payments. Self-hosted. Built on Claude. | TypeScript | +627 | 2,102 |
| 08 | nashsu/llm_wikiLLM Wiki is a cross-platform desktop application that turns your documents into an organized, interlinked knowledge base — automatically. Instead of traditional RAG (retrieve-and-answer from scratch every time), the LLM incrementally builds and maintains a persistent wiki from your sources。 | TypeScript | +640 | 18,683 |
| 09 | sponsors/obraAn agentic skills framework & software development methodology that works. | Shell | +731 | 285,331 |
| 10 | Sonarr/SonarrSmart PVR for newsgroup and bittorrent users. | C# | +174 | 15,714 |
02 / MODELS
| 排名 | 模型 | 任务 | 收藏 | 下载 |
|---|---|---|---|---|
| 01 | Z-Image-Turbo | 文本生成图片 | 2,496 | 803,103 |
| 02 | MiniMax-H3 | — | 1,963 | 137,632 |
| 03 | Qwen3.8-27B | 视觉多模态理解 | 1,405 | 266,602 |
| 04 | DeepSeek-R1 | — | 1,381 | 927,318 |
| 05 | Qwen3.8-Flash-Next | 视觉多模态理解 | 1,148 | 48,004 |
| 06 | Qwen3-Coder-480B-A35B-Instruct | 文本生成 | 1,093 | 138,497 |
| 07 | IndexTTS-2 | 语音合成 | 975 | 207,349 |
| 08 | chatglm3-6b | — | 830 | 1,590,754 |
| 09 | Qwen3.8-2.4T-A95B | 文本生成 | 759 | 1,858 |
| 10 | Qwen-7B-Chat | 文本生成 | 757 | 693,248 |
DEVELOPER VIEW
值得关注的理由各不相同。i-have-adhd 和 obra 几乎零成本可试，改的是日常与 agent 协作的体感，投入产出比最高。PI-Desktop 用 Rust 做宿主核心，对在意本地推理、插件生态和长期可维护性的团队有参考价值。llm_wiki 提供了一条与向量库路线不同的知识库架构，适合文档量大、希望复用已有语料的场景。CloddsBot 与 DeskcommCRM 属于高风险高信息量项目——前者涉及真实交易执行，后者涉及合规，读代码比直接部署更有价值。模型侧 Qwen Coder 480B-A35B 与 Flash-Next 分别代表大而专和快而广两条选型路线，适合作为基座评估起点。
WORTH A STAR
1. gods-eye-view：把真实开源空间数据渲染进写实 3D 地球，浏览器内即可运行，是"数据可视化 + 地理情报"少见的完成度组合。
2. llm_wiki：用增量构建的持久维基替代每次重跑的 RAG，是知识库架构上一个值得验证的新范式。
3. i-have-adhd：改动极小但直击痛点，是"agent 输出工程"最容易立刻落地的一个。
SOURCES & METHOD
GitHub Trending（项目热度、语言、Star） · ModelScope（模型任务、收藏与下载）
抓取日期：2026-09-12。排名是当日快照，指标可能随平台更新而变化。