Case StudyFounder, Architect & AI Engineer

Docutron.

An AI document-intelligence SaaS that turns raw project documents into compliant, investor-grade technical reports using a 15-agent Claude pipeline with strict anti-hallucination sourcing.

Client
Independent Product
Year
2026
Tools
Claude API, FastAPI, Python, PostgreSQL, Celery, Redis, React, TypeScript, Stripe

Overview

Docutron tackles a genuinely hard document problem: taking a pile of raw project files — drill logs, permits, assay results, surveys, photos — and producing a compliant, investor-grade technical report. The founding use case is the NI 43-101 mining standard, where accuracy is a regulatory requirement, not a nicety. The platform's core principle is anti-hallucination by construction: every AI-generated claim must cite a source document and page reference, and missing data produces an explicit gap statement rather than a guess.

The Multi-Agent Pipeline

Fifteen specialized Claude agents collaborate on each report: an orchestrator coordinates ingestion, classification, section extraction, QA, editorial review, economic analysis, permit and timeline agents, image handling, and a final report builder. A schema-discovery agent researches regulatory standards using web search, so the system can adapt to new report formats. Model selection is tiered — Haiku for high-volume classification, Sonnet for extraction and writing, Opus where depth matters.

Engineering Highlights

  • Prompt caching on agent system prompts cuts AI costs by more than 80% across the pipeline
  • Per-call AI usage tracking, cost estimation, and circuit breakers around every Claude integration
  • Async FastAPI backend with SQLAlchemy 2.0, Celery/Redis task queues, and schema-per-tenant PostgreSQL multi-tenancy
  • Security-first auth: JWT RS256, bcrypt, and TOTP two-factor authentication
  • Stripe billing across three subscription tiers, with Whisper-powered voice transcription for interview ingestion

Outcome

A deeply architected demonstration of production multi-agent design — orchestration, cost control, resilience, and verifiable output — applied to a domain where hallucination is disqualifying.

Key Outcomes

  • 15-agent Claude pipeline: orchestration, ingestion, classification, extraction, QA, editorial, economic, and report-building agents
  • Strict anti-hallucination design — every AI claim cites a source document and page, with explicit gap statements for missing data
  • Prompt caching on agent system prompts for 80%+ AI cost savings, plus per-call usage tracking and circuit breakers
  • Schema-per-tenant multi-tenancy on PostgreSQL with JWT auth, TOTP 2FA, and Stripe subscription tiers
  • Founding use case: NI 43-101 mining technical reports assembled from drill logs, permits, assays, and surveys