Ennis Studio

Work

Systems I build and run myself.

Client work stays confidential. These are my own products, so I can show you exactly how they work and what went wrong along the way.

CS-01 · Data platform

CertIndex

ctindex.io ↗Infographic: CertIndex pipeline from public Certificate Transparency logs through ingest, deduplication and indexing to a REST API, live event stream and MCP server used by developers, security tools, AI agents and platforms

A real-time index of Certificate Transparency logs, served to developers through a REST API, a live event stream and an MCP server that AI agents query directly.

Problem
Public CT logs record every TLS certificate issued on the web, but the raw logs are huge, append-only and awkward to search. Security teams and tools need them indexed, deduplicated and queryable in seconds.
Built
  • Continuous ingestion from public CT logs into an indexed store
  • REST API with search by domain, issuer and time window
  • Server-sent event stream for new certificates as they appear
  • MCP server so AI agents can look up certificate data as a tool
  • Self-serve plans plus a licensing path for platforms that redistribute the data
Lesson
Running a data pipeline in production teaches you where AI projects actually fail: data freshness, cost of storage, rate limits and observability. Those lessons go straight into Build Days and due-diligence reviews.
Related
AI Build Day · Technical Due Diligence

CS-02 · Agent loop · Editorial

ChangeRiskIntel

changeriskintel.com ↗Infographic: ChangeRiskIntel editorial agent loop of schedule, research, draft, human approval gate, publish and recheck, covering five desks

An editorial site on IT change risk, operational risk and security operations, researched, published and maintained by scheduled AI agent loops.

Problem
Specialist content needs research, sourcing, editing and upkeep. Doing that by hand for a niche audience rarely pays. The question was whether an agent loop could do it to a standard worth reading.
Built
  • Scheduled research and drafting runs with source requirements
  • Repository-committed publishing, so every change is reviewable
  • Automated freshness checks on existing pages
  • Human approval gates for anything irreversible
Lesson
Agent loops work when the rules are written down and every output is checked. That is the same discipline the Safe AI Adoption Workshop puts into your policy.
Related
Safe AI Adoption · Market Research Sprint

CS-03 · Agent loop · Tools

Control Plane Labs

controlplanelabs.com ↗Infographic: Control Plane Labs offers 21 free browser tools kept current by a scheduled loop of sources, code review and publishing

Free browser-based tools and practical reference guides for DevOps, security and web development, kept current by the same kind of scheduled loop.

Problem
Tool and reference sites go stale fast. Keeping dozens of pages accurate is maintenance work that people stop doing.
Built
  • Browser-based utilities that run without a backend
  • Reference pages generated and updated from cited sources
  • Scheduled code review and issue filing against the site repo
Lesson
The most useful automation is often maintenance, not creation. Many Opportunity Workshops end with a maintenance task ranked first.
Related
AI Opportunity Workshop

Want this kind of system in your business?

Book a free 30-minute fit call. Answer three questions first so the call is useful, then pick a time on my calendar.

Book a fit call Take the scorecard