LokaAudit is an advanced security auditing platform that combines automated vulnerability detection with Machine Learning powered analysis to deliver comprehensive, enterprise-ready audit reports.
AuditFi is an AI-powered, multi-chain smart-contract auditing platform designed to solve one of the biggest problems in Web3 today: slow, expensive, and EVM-only audits. Traditional security audits take weeks, cost anywhere from $1,000 to over $100,000, and still leave massive gaps—especially for non-EVM chains like Solana, Aptos, and Sui, where automated auditing tools are almost nonexistent.
AuditFi changes this landscape completely.
It combines LLM-based static analysis, dynamic test-case generation, virtual execution sandboxes, and on-chain verification to produce actionable security reports within minutes. The system supports Solidity, Rust, and Move smart contracts, enabling developers from multiple ecosystems to get instant feedback and fix vulnerabilities early.
AuditFi democratizes blockchain security by making enterprise-grade audits accessible to solo developers, hackathon teams, startups, and DAOs. With real-time testing, auto documentation, and multi-chain support, AuditFi acts like:
ChatGPT × CertiK × Jenkins for Smart Contracts — automated, fast, affordable, and chain-agnostic.
It aims to reduce ecosystem-wide vulnerabilities, prevent costly exploits, and help teams ship safer Web3 products with confidence.
During the hackathon, we achieved significant progress across the platform’s core features and architecture: ✅ 1. Functional Multi-Chain Code Parsing Implemented initial parsing pipelines for Solidity, Rust (Solana), and Move (Aptos/Sui). Standardized code abstraction layer for unified audit processing. ✅ 2. AI-Powered Static Vulnerability Analysis Integrated LLaMA 3.1, DeepSeek-Coder 70B, and Gemini 2.5 for hybrid analysis. Built prompts and agents for identifying common vulnerability patterns (reentrancy, overflow, access control issues, unbounded loops, etc.). ✅ 3. Dynamic Testing & Auto Test-Case Generation Implemented automated test-case generation for detecting runtime vulnerabilities. Created an execution sandbox for safely running untrusted contract logic. ✅ 4. Report Generation System Built the first version of automated audit reports, including: vulnerability summaries severity scoring mitigation suggestions auto-generated documentation ✅ 5. Dashboard + UI Prototype Developed the Next.js + Tailwind front-end. Added code upload, analysis status, vulnerability display, and report download. ✅ 6. Backend & API Integration Set up FastAPI + Node.js backend structure. Connected LLM inference endpoints and audit workflow engine. Added MongoDB for storing analysis logs, contracts, and reports. ✅ 7. Containerization & Deployment Dockerized backend and inference modules. Initial deployment tested on Render/Fly.io. Preparing production-grade decentralized deployment on Akas Network (Cosmos). ⚠️ 8. Challenges Identified A major challenge is the lack of openly available datasets for non-EVM audit reports, making supervised model training difficult. We designed a strategy using synthetic datasets, fuzzing traces, and open-source contract corpora to mitigate this.