polybrain.ethv0 • experimental
Doxel
Agent-Native Project
29 ptsScore (v2)
This grant is under review.
Evaluation VersionCurrent Evaluation (o3)
AI Evaluation Feedback (v2)
Doxel aims to become an open-source, Polygon-integrated document AI agent that transforms unstructured PDFs into queryable intelligence, targeting legal, healthcare and DAO governance use-cases. The large and fast-growing IDP market and the team’s verified software engineering internships provide a credible foundation, while an OSS distribution strategy could harness developer momentum on GitHub. However, differentiation from well-funded incumbents, a name collision with construction-tech firm Doxel.ai, absence of a public repo and lack of quantified user/adoption targets leave execution risk. Strengthen the application by releasing an initial GitHub prototype, outlining a smart-contract architecture for Polygon fee flows, securing at least one domain expert or community advisor as a voucher, and providing concrete OKRs (e.g., 1 k GitHub stars, 50 on-chain document proofs within 6 months). With these improvements and clearer POS impact, the project could merit a full 40 k POL grant; at present it is a promising but early bet.
FoundationScore: 6/10
Automates ingestion, semantic search and Q&A over unstructured documents for use-cases such as contract analysis and policy transparency.
Team ExperienceScore: 6/10
Hamza Dincer – Junior software developer (Zyfera) with active GitHub portfolio; Halit Dincer – CS student & ex-Software Eng. Intern at Mark43 with personal site and GitHub.
Distribution StrategyScore: 5/10
Launch OSS repo + HuggingFace demo, cross-posted on X, HackerNews and Reddit; plugin SDK to spur community extensions.
Milestones & BudgetScore: 6/10
M1: parsing & base RAG; M2: structured extraction; M3: multi-doc agent – each 1 month.
Technical IntegrationScore: 5/10
Future hosted API could settle usage fees in POL and write document provenance hashes on-chain.
Community Feedback (Bonus)Score: 0/5
No external vouches submitted.
Project Details
Target AudienceLegal teams, healthcare payers, DAOs and other knowledge-intensive Web2/Web3 organizations that handle large PDF/Doc archives.
Success MetricsEarly goal is a functional OSS agent in 3 months; long-term metrics would be GitHub stars, monthly active API calls and on-chain usage of Polygon-integrated features.
Budget Request10 k POL per milestone = 30 k POL total (~$18 k at 0.60 $POL). No allocation for marketing or audits.
Open Source Bonus
+1 points
Project pledges full MIT-licensed repo and HuggingFace demo but repo is not yet public.