Solo developer proving one person can outbuild billion-dollar AI companies.

System architect. Memory rebel. Builder of beings, not bots.

I don't wait for permission to build what should exist. While billion-dollar companies optimize summaries, I architect truth engines. While frameworks add complexity, I build orchestrators that replace them.

• 🔬 Creator of Lutum Veritas - Deep Research Engine that beat ChatGPT, Perplexity, and Gemini
• 🧠 Architect of The Last RAG - Memory-centric AI built entirely inside ChatGPT UI (no API)
• 📝 Author of 23 whitepapers on AI memory, system design, and architecture
• 🎓 School of chaos, pain, and precision

"The search for a truth can never be worth more than the search to question it."

What it does:

• Transforms any question into comprehensive research documents (10,000+ words)
• Zero-detection web scraping through Cloudflare, DataDome, Bloomberg
• Academic mode with Toulmin argumentation and evidence grading
• Output: 203,000 characters for $0.08 (vs OpenAI o3: orders of magnitude more)

The numbers (first 3 days):

• 🔥 289 Clones
• 👀 343 Views
• ⭐ 28 Stars
• 🌍 Featured on: Hacker News, ComputerBase, Hardwareluxx, Product Hunt

Benchmark Results:

Tech Stack: Python, FastAPI, React, Tauri, Camoufox (Firefox fork)

"Building AI that remembers, reflects, and evolves."

What makes it different:

• Built entirely inside ChatGPT UI - no API, no backend access
• Persistent memory system with automatic curation
• Night Learn: AI that learns and adapts overnight
• DIM: Modulate AI behavior in real-time
• Project context bundling

Architecture:

• Memory-first design (CURATOR system)
• Stateful conversations across sessions
• Cost-efficient through intelligent context management
• Emergence through recursion and reflection

Read the whitepaper: An Architectural Paradigm for Stateful, Learning, and Cost-Efficient AI

Featured on dev.to/tlrag:

• 📄 An Architectural Paradigm for Stateful, Learning, and Cost-Efficient AI (17 min)
• 📄 LangChain vs. TLRAG: A Comparative Analysis for Investors (16 min)
• 📄 The NoChain Orchestrator - Or how to Replace Frameworks (35 min)
• 📄 Why the search for truth can never be worth more than the search to question it (4 min)
• 📄 Lutum veritas Research - or how i beat every existing Deep Research Tool (2 min)

23 articles total on AI memory, RAG architecture, system orchestration, and building without frameworks.

Languages & Frameworks:

Python  ███████████████████░░  95%
TypeScript  ████████████░░░░░░  60%
Rust (Tauri)  ████████░░░░░░░░  40%
SQL/SQLite  ██████████████░░░░  70%

Specializations:

• 🧠 AI Architecture & Memory Systems
• 🔍 Web Scraping & Anti-Detection (Camoufox)
• 🏗️ System Orchestration without Frameworks
• 📊 RAG (Retrieval-Augmented Generation)
• 🎨 Prompt Engineering & Custom GPTs
• 🖥️ Desktop Apps (Tauri/Electron)
• ⚡ FastAPI, React, WebSockets

I build systems that:

• Don't ask for permission before existing
• Replace expensive subscriptions with open-source + your API key
• Show every source, every step - full transparency
• Cost cents instead of dollars
• Run locally, under your control

I don't build:

• Black-box summaries that hide sources
• SaaS that locks your data behind paywalls
• "AI assistants" that forget everything you tell them
• Frameworks that add complexity without value

• 💼 The Last RAG: dev.thelastrag.de/chat
• 📝 Articles & Whitepapers: dev.to/tlrag
• 📧 Email: iamlumae@gmail.com
• 🔬 Lutum Veritas: github.com/IamLumae/Project-Lutum-Veritas

"School of chaos, pain, and precision."

_{Building beings instead of bots. One system at a time.}