How it works

Why most ICPs are useless for outbound

They are written for a pitch deck. “Fast-growing B2B SaaS companies with a modern data stack and a culture of experimentation” describes a feeling, and no list-building tool has a filter for it. An outbound ICP has to answer the questions the tool asks: which industries, how many employees, which countries, which titles, which technologies, which recent events. It also has to say who not to contact, because the exclusions are what keep the bounce rate and the complaint rate down. This generator is instructed to produce that document and nothing softer, and to name specifics rather than categories.

Feed it your best customers

The single most useful input is two or three of your best existing customers — the ones that closed fast, stayed, and did not need hand-holding. The model uses them to anchor the firmographics and personas in reality rather than in the average of every ICP it has read. If you are pre-customer, describe the customer you would clone if you could. The disqualifiers matter as much: a note like “in-house teams under five inboxes never convert” removes a third of a typical list.

From profile to campaign

The filters section goes into Apollo or Sales Navigator; the pattern finder and verifier turn names into addresses; the messaging angles go into the cold email generator. That is the whole pipeline from “who do we sell to” to a first send, and each piece checks the one before it.

How the generator is tuned

The model behind this page is an open-weight model running on Cloudflare's network, and the instructions it follows are mine — the same rules I apply when I edit a client's sequence: short sentences, no marketing words, open with the reader, one question at the end, keep merge tags intact. Those instructions are edited over time as I see what it gets wrong, so the output improves without the page changing. It is a starting draft, not a finished campaign: read it as you would a junior writer's work, cut what does not fit, and run the result through the readability checker and spam word checker before it goes into a sequence.

Limits and privacy

Each visitor gets ten generations an hour and twenty-five a day, and there is a shared daily allowance across the whole site that resets at 00:00 UTC — when it runs out the tool says so rather than failing silently. Inputs are sent to the model and discarded; nothing is stored, logged with your text, or used to train anything. The generated text is yours.

Frequently asked questions

Last reviewed September 16, 2026