About

ExpLint is an MCP plugin that connects to agents like Claude, ChatGPT, and Grok Bot to check A/B test specifications for errors and flawed logic.

Where did you start?

Kameron Tanseli, the founder of ExpLint, has run over 1,500 experiments across his nine-year career as a growth engineer. In the last year at Fyxer, where he was the founding growth hire, he launched 540 experiments while the company grew from $1M to $35M ARR in 13 months.

ExpLint was born from the need for one consistent, reliable way for everyone in the company to check that an A/B test idea was properly planned and followed best practices before it entered the software factory, where AI writes the code and ships it.

Who are you doing it for?

If you run experiments and want your agents to take over writing and planning them, ExpLint is built for you.

Growth engineers, product managers, and CRO consultants rushing between tasks get a dependable linter that catches their errors, which is the safety net that keeps an experimentation programme high quality.

Why you and why now?

At FLORA, the Growth Engineering team ships 6–8 experiments a week across the product. The team is exactly one person: me.

That velocity is possible because of a semi-autonomous software factory built with Devin, which drafts the ticket with help from ExpLint, designs the treatment UI, builds it, then sets up the PostHog experiment and launches it to users.

Where are you going?

First, getting agents to write and plan experiments properly and consistently.

Next, reviewing and improving treatment UIs with vision models and growth tactics.

Finally, turning finished specifications into a memory and knowledge base that fills itself in and starts suggesting ideas for your experimentation programme.