{
  "id": 14329203,
  "title": "Building LexAegis: How Qwen 3.8 Max Powers an Evidence-Backed Legal AI Assistant on Monad",
  "url": "https://urgent.news/2026/10/14/building-lexaegis-how-qwen-3-8-max-powers-an-evidence-backed-legal-ai",
  "topic": "ai",
  "section": "AI",
  "published": "2026-10-14T00:26:40.000Z",
  "source": {
    "name": "Dev.to",
    "slug": "dev-to",
    "url": "https://dev.to/eni_agunbiade_20da61cf0bf/building-lexaegis-how-qwen-38-max-powers-an-evidence-backed-legal-ai-assistant-on-monad-455l"
  },
  "original_language": "en",
  "account": "Creating LexAegis: How Qwen 3.8 Max Powers an Evidence-Backed Legal AI Assistant on Monad\n\nThe goal was to build an AI-assisted legal workspace that separates AI-assisted analysis from evidence verification and human approval. The Monad Metropolis hackathon project, LexAegis, combines Qwen 3.8 Max, encrypted document storage, human-controlled review, and blockchain-based Proof of Review.\n\nQwen 3.8 Max serves as the reasoning and language component. Integrated through Alibaba Cloud's compatible-mode API, the model receives structured JSON output and native tool calling. LexAegis gives Qwen a constrained workflow rather than unrestricted legal authority.\n\nLexAegis implements four read-only assistant tools:\n1. read_sources - retrieves evidence excerpts for the model's responses\n2. check_payments - examines payment figures using deterministic code\n3. check_evidence - identifies evidence gaps\n4. get_status - retrieves application or review status information\n\nThe model can request these tools, but the application validates arguments and controls execution. Tools have limits on execution time, provider attempts, and usage. The model cannot use these tools for legal actions like signing agreements or altering confirmed facts.\n\nA test using a fictional employment settlement agreement demonstrated Qwen's ability to detect discrepancies. Qwen identified a £2,000 difference between stated and instaled amounts, incorporating the calculation into its response. The deterministic payment checker confirmed the £2,000 discrepancy, highlighting the importance of combining AI interpretation with deterministic calculations and verification.\n\nLexAegis' conversational interface allows users to ask Aegis to analyze legal matters, retrieve relevant excerpts, and provide evidence-backed responses. The conversational Command Centre was deployed on Railway on October 13, 2026. Privacy is maintained through client-side encrypted storage and an explicit disclosure process before material is sent to Qwen.",
  "summary": "Building LexAegis: How Qwen 3.8 Max Powers an Evidence-Backed Legal AI Assistant on Monad Building AI that helps professionals make decisions is one challenge. Building AI whose evidence, limitations and human approvals can be inspected is another. For the Monad Metropolis hackathon, I built LexAegis, an AI-assisted legal workspace that combines Qwen 3.8 Max, encrypted document storage,…",
  "key_points": [
    "LexAegis combines Qwen 3.8 Max with encrypted document storage and blockchain-based Proof of Review",
    "Four read-only assistant tools: readsources, checkpayments, checkevidence, and getstatus",
    "Qwen detected £2,000 discrepancy in fictional employment settlement agreement"
  ],
  "editors_take": "Integrating reasoning models with deterministic tools and blockchain verification shifts legal AI from speculative generation toward a structured framework of evidence-backed analysis and human-controlled oversight.",
  "illustration": null,
  "coverage": {
    "outlets": 1,
    "also_reported_by": []
  },
  "ai_generated": true,
  "disclaimer": "Summaries, key points and the editor’s take are written by software from other outlets’ reporting and may contain errors — always check the linked original."
}