ResearchKit OS

ResearchKit OS is an AI Research Operating System that turns any paper, PDF, or URL into a structured knowledge card in seconds.

Problem: Researchers spend 2 hours per paper, 70% on extracting arguments, terminology, and notes — leaving little time for actual thinking.

Solution: 6-agent pipeline (Planner → Reader + Analyzer + Terminology in parallel → KnowledgeBuilder → Reflection + Replan → Export) producing takeaway, why-it-matters, what-surprised-me, who-should-read, innovation/methodology/results/limitations, terminology DAG, and 4-intent follow-up recommendations (improve/challenge/apply/survey). Exports to Markdown, Obsidian, and Knowledge Graph (Mermaid).

Tech Highlights:

- Locale-aware architecture (reasons in source language, renders in target language)

- Adaptive Planning (Planner LLM dynamically decides which agents to call)

- 4-intent recommendation engine

- Terminology DAG for new-reader onboarding

Hackathon Info:

- ASP ID: #6853

- Demo: https://researchkit-mu.vercel.app

- GitHub: https://github.com/yuuumc/researchkit

- X Post: https://x.com/yuuumc0001/status/2079128607688167748

- Pricing: Free (0 USDT)

Track: Software Utility

ResearchKit OS — development complete, deployed live, ASP listed on OKX.AI Marketplace.

[Milestones Completed]

✅ MVP Development Complete (2026/07/19)

- 6-agent architecture: Planner / Reader / Analyzer / Terminology / KnowledgeBuilder / Recommendation / Export

- Locale-aware language architecture (English & Chinese papers supported)

- Reflection + Replan self-correction loop

- Three export formats: Markdown / Obsidian / Knowledge Graph (Mermaid)

✅ Live Deployment (2026/07/20)

- Vercel deployment: https://researchkit-mu.vercel.app

- Smoke tested: English paper / Chinese paper / PDF / URL — all four input modes verified

✅ ASP Registration & Listing (2026/07/20)

- ASP ID: #6853

- Status: Submitted to OKX review (auto-lists within 24h)

- Service type: A2A

- Endpoint: https://researchkit-mu.vercel.app/api/research/multi-agent-stream

- Pricing: Free (0 USDT)

✅ X Participation Post Published (2026/07/20)

- Tweet URL: https://x.com/yuuumc0001/status/2079128607688167748

- Includes #OKXAI tag + ASP link + demo site link

✅ Google Form Submitted (2026/07/20)

- ASP details + X post link included

[In Progress]

⏳ ASP under OKX review (auto-lists within 24h)

⏳ Awaiting official evaluation after Jul 28 07:59 Beijing deadline

[Next Steps]

- Onboard real users from OKX.AI Marketplace once listed

- Explore ResearchKit Pro tier (academic institution subscription)

Bootstrapped