Useful things, free.
No email needed. No signing up. Just run them in your browser, print the result, and bring it to your board. If they help, great. If not, ignore them.
Time-savings calculator
Three sliders, honest numbers. See how much time your team could save if software handled the busywork.
Open the calculator →
AI readiness quiz
Eight quick questions. Tells you honestly whether AI is the right next step for your organization — or whether something simpler comes first.
Take the quiz →
Tools we built, released for anyone.
Small utilities that came out of real client and product work. MIT licensed, no strings — the code is yours to read, run, and change.
RubySage
Chat with your Rails codebase. Scan your app, ask questions, get answers grounded in your actual code — with file citations.
View on GitHub →
Agent Eval Harness
Vendor-neutral regression testing for AI agents. Versioned eval cases, deterministic checks plus an LLM-as-judge, and CI that fails when behavior drifts.
View on GitHub →
retrospec
Mine your repo's git history into a private eval suite for AI code models. Every bug fix that also changed a test becomes a case the harness can grade. One file, no dependencies, nothing leaves your machine.
View on GitHub →
Permissioned MCP Server
A reference MCP server showing what the quickstarts skip: least-privilege tools, execution-time authorization, destructive-action confirmations, and metadata-only audit logs.
View on GitHub →
tapewarp
Make any video look like it was shot on a 1990s camcorder and played back off a worn VHS tape. One Python file, no ML, no cloud, no watermark.
View on GitHub →
saywhat
Get the transcript of any video, locally. Platform captions first (YouTube, TikTok, ~1800 sites), local whisper transcription when there are none. One Python file, no API keys.
View on GitHub →
localflight
Build your Expo iOS app on your own Mac and submit straight to TestFlight — zero EAS cloud build credits. Auto-detects bundle ID, build number, and App Store Connect ID from your project files.
View on GitHub →
cutshort
Turn long video into captioned vertical shorts — ffmpeg + local whisper, word-synced highlight captions, 9:16 reframing, music beds. The engine behind a nightly shorts pipeline, minus the branding.
View on GitHub →
Tools tell you where to start. A conversation tells you how.
Once you have a number or a score, the next thing is figuring out what to do with it. Fifteen minutes is usually enough to get you started.
Book a free 15-min call