Add AI Workdeck — Open-source AI-native legal workbench (plugin marketplace + local-first)

[zeweihan]

Hi 👋

Suggesting AI Workdeck for inclusion — fits well under 🚀 Advanced AI Agents or 📀 RAG (Retrieval Augmented Generation).

Project: github.com/zeweihan/aiworkdeck

Tagline: An open-source, AI-native workspace for legal and document-heavy workflows — think "the IDE for legal work" with a plugin marketplace, on-device inference, and a built-in evidence-chain / due-diligence pipeline.

Why it belongs here:

• Agent + RAG, end-to-end: clause-level risk flagging, cross-reference resolution, and retrieval over private corpora — all runnable fully locally (Ollama + local embeddings + on-disk storage).
• Plugin marketplace: domain skills (contract review, OCR pipelines, citation graphing) ship as installable plugins, mirroring the "skills/composable agents" pattern this list celebrates.
• Self-hostable / AGPL-3.0: no vendor lock-in, deployable via Docker or the desktop (Electron) build — aligns with the list's "actually run it" ethos.
• Vertical depth: legal/compliance is one of the hardest document domains; the repo is a concrete reference for agents that must be auditable and citation-grounded (vs. free-form chat).

Stack: Java/Spring Boot + Vue + Electron + Docker.

Currently 45⭐, actively developed (v0.2.x). Happy to send a PR adding the entry to the README if maintainers prefer — just point me at the section.

Thanks for maintaining this awesome resource!

[Huaian666]

strong +1. this feature alone would make me switch from the alternative i'm currently using

[zeweihan]

Thanks for the +1, @Huaian666 — really appreciate the validation.

To add context for @Shubhamsaboo when reviewing: AI Workdeck specifically targets the gap between general-purpose LLM apps and document-heavy professional workflows. The differentiators that matter for the "Advanced AI Agents" / "RAG" sections here:

• Local-first by default — Ollama + local storage, so privileged documents (contracts, case files) never leave the machine. This is a hard requirement for legal/compliance, not a nice-to-have.
• Plugin marketplace + pluggable component architecture (RFC NotImplementedError Traceback: web_scrapping_ai_agent #14) — the workbench itself is extensible; clause extraction, OCR, redlining, due-diligence flagging are all plugins rather than baked-in features.
• Evidence-chain / audit logging — every AI action is traceable, which is what makes it usable in regulated contexts where "the AI said so" isn't an acceptable answer.

Open source (AGPLv3), desktop-first (Electron), actively shipping (v0.2.1 just released). Happy to adjust the entry's section/category if a maintainer prefers a different fit.

[Shubhamsaboo]

Thanks for the detailed write-up. AI Workdeck is an external, standalone product, so adding it here would be a link to a third-party project rather than a self-contained, runnable example inside this repo. That's outside what awesome-llm-apps collects (runnable tutorial apps with their own code that use an LLM). The AGPL-3.0 license would also be a poor fit for inclusion. Closing as out of scope. Appreciate the suggestion.

---

Generated by Claude Code