Articles and guides
Blog
Hands-on notes from building private AI, automations, and safety controls for real businesses. 26 articles.
All 78 topics
Featured Preventing Prompt Injection in Local LLMs: A Working Defense
How to detect and block prompt injection attacks in self-hosted LLM deployments using a dedicated classifier model in Open WebUI — with a nine-category attack taxonomy, violation logging, and lockout.
Featured Implementing ClamAV File Scanning in Open WebUI
Step-by-step guide to adding real-time antivirus scanning to Open WebUI file uploads with ClamAV: Docker setup, filter installation, violation logging, and automatic user lockout.
Open WebUI Knowledge Bases: RAG Made Simple for Real Teams
A practical guide to Open WebUI Knowledge Bases and beginner-friendly RAG workflows for businesses that need grounded answers from their own documents.
HR And Benefits Teams Need Private AI Before Employee Documents Go Into Chatbots
A practical private AI use-case guide for HR and benefits teams handling employee documents, policies, investigations, and sensitive internal records.
When Your Brand's AI Answers Python Questions: Why Scope Control Needs LoRA-Tuned Open-Weight Models
Why off-topic AI answers are a governance failure, and how LoRA fine-tuning on open-weight models gives businesses real control over what their assistant will and will not do.
Accounting Firms Should Not Put Client Tax Packets Into Public AI
How accounting and advisory firms can use private AI during busy season without turning client tax and financial documents into unmanaged uploads.
Open WebUI Is The Free And Private ChatGPT Alternative Businesses Should Know
A video-backed guide to Open WebUI as a private ChatGPT alternative for teams that need control over models, documents, users, and data.
Health Insurance AI Has To Start With Member, Claims, And Document Boundaries
A practical guide to private AI use cases for health insurance teams that need document support without uncontrolled member or claims exposure.
Corporate Insurance Teams Need Private AI For Submissions, Claims, And Renewal Work
How corporate insurance teams can use private AI for submission review, claims support, and renewals without exposing sensitive client documents.
Why Small Law Firms Need Private AI Before They Need More Software
Small law firms do not need another generic AI app first. They need a governed private AI workspace for confidential legal documents.
Small Law Firms Do Not Need More AI Hype. They Need A Safe Document Workflow
A practical private AI starting point for small legal firms that need useful document workflows without unmanaged client-data exposure.
Your Employees Are Already Using AI With Client Documents. Now What?
A practical guide for small firms that need to turn shadow AI use into a governed private AI workspace before sensitive documents leak.
Shadow AI Is Already Inside Your Business. Here Is How To Bring It Under Control
A small-business guide to turning unmanaged employee AI use into a governed private AI workspace with practical controls.
What It Means for a Website to Be Agent-Native
A practical breakdown of how damore.ai became an interactive, machine-verifiable AI site with MCP, WebMCP, DNS-AID, and agent discovery.
Training Distinct LLM Voices and Orchestrating Them in a Live Multi-Speaker Pipeline
A technical walkthrough of two open-source projects: a QLoRA fine-tune pipeline that trains 26 distinct personas onto Qwen3-14B, and the Open WebUI pipeline that turns those voices into an orchestrated multi-speaker conversation.
Air-Gapped AI for Government and Defense: Deploying LLMs in Disconnected Environments
How to deploy AI in fully air-gapped or SCIF environments — no internet, no cloud, no external dependencies. Architecture, model selection, and operational considerations for government and defense teams.
HIPAA-Compliant AI Deployment: What Healthcare IT Leaders Need to Know
A practical guide to deploying large language models in healthcare environments while maintaining HIPAA compliance — covering architecture, data isolation, audit logging, and vendor risk.
Private LLM Deployment: Running AI On-Premise Without the Cloud
How to deploy large language models entirely on your own infrastructure — no API keys, no cloud dependencies, no data leaving your network. A guide for IT leaders evaluating on-premise AI.
Omnichannel Private AI: OpenClaw + OpenWebUI Architecture
How to route WhatsApp, Telegram, Discord, Signal, and Slack through your private AI stack — with safety filters, persistent memory, RAG, and full audit logging on every channel.
QLoRA vs Full Fine-Tuning: When Does Precision Matter?
A practical guide to choosing between QLoRA and full-precision fine-tuning for domain-specific LLMs. When 4-bit training is the right default — and when it's not.
Building a Clinical Oncology AI: From PubMed Abstracts to a Thinking Medical Model
A deep dive into the end-to-end pipeline for fine-tuning Qwen3 14B into a clinical oncology reasoning model — from synthetic data generation with anti-hallucination safeguards to two-phase SFT + DPO training on NVIDIA DGX Spark.
Unsloth: Training Bigger Models on Smaller Hardware
How Unsloth's pre-quantized models and optimized training toolkit let you fine-tune larger, more capable LLMs on resource-constrained hardware — with a real-world case study on the Cross & Faith AI platform.
Token Usage Tracking & Rate Limiting for Open WebUI
A group-based token usage tracking and rate limiting filter for Open WebUI with PostgreSQL persistence, automatic schema creation, and zero PII storage.
PubMed Deep Research Tool: AI-Powered Medical Literature at Your Fingertips
Introducing the PubMed Deep Research Tool for OpenWebUI — search, deduplicate, and archive PubMed articles into a growing knowledge base with NLP-powered entity extraction and intelligent query optimization.
Announcing the Dynamic Circle of Speakers v2.0
Introducing a first-of-its-kind multi-speaker AI pipeline that brings together a curated group of distinct voices to respond to your questions in real time.
Three Layers of Safety: How We Protect Every Conversation
Every conversation on the Cross & Faith AI platform passes through three independent safety filters — prompt injection protection, content safety, and policy enforcement.
Let's find your first three wins
A free 20-minute call. Bring the workflow that annoys you most. I will tell you honestly whether AI fixes it and what the first stage would cost.
Prefer to write first? Send a short inquiry and I reply within one business day.