Blog
Product notes, AI automation experiments, comparisons, and implementation learnings behind the products Carlos builds at IliciLabs.
Aurora Subtitles
Popular Aurora guides
Useful reads for Discord live captions, hearing support, local-first processing, and the difference between text translation and live speech translation.
How to Choose the First Business Process to Automate with AI
A practical framework for choosing the first AI automation project without overbuilding or increasing operational risk.
Human-in-the-Loop AI Agents: A Practical Checklist for Companies
Checklist for designing AI agents with review points, logs, escalation paths and clear human responsibility.
Local-First AI Automation: When Company Data Should Stay on Your Machine
How to decide when local-first AI automation is better for sensitive company data, workflows and compliance.
Private Meeting Notes with AI: Summaries Without Losing Control
How to use AI for private meeting notes, summaries and follow-ups while keeping human control and sensitive data boundaries.
How to Make a Product Website Useful for AI Agents and LLM Search
A practical guide to making product websites understandable for AI agents, LLM search, Google, and human buyers.
Real-Time Captions for Family Video Calls and Caregiving on Windows
How real-time captions can help family calls, seniors, caregiving contexts and hearing accessibility on Windows.
Text Translation vs Live Speech Translation on Windows: Which Tool Fits Your Workflow?
Compare text translation and live speech translation on Windows, and when LuminaL or Aurora Subtitles fits better.
Vision Accessibility with AI: OCR, Scene Description, and Desktop Workflows
How AI can support low-vision and vision accessibility workflows through OCR, reading assistance and local desktop tools.
Agent Workflow Architecture for Business Processes
A practical architecture for AI agent workflows: triggers, tools, state, permissions, logs, review queues, and fallback paths.
AI Agent Governance: Logs, Permissions, and Human Checkpoints
Useful AI agents need governance: clear permissions, audit logs, review queues, escalation paths, and human checkpoints where risk matters.
AI Agents in Business Processes: Keep Humans in Control
AI agents can coordinate business workflows, but useful automation needs human checkpoints, auditability, and clear boundaries before autonomy.
AI Automation Data Readiness: Fix the Inputs Before the Agent
AI automation depends on clean inputs, clear sources of truth, access rules, and operational context. Fix those before blaming the model.