One person, 17 data sources, and a strong opinion about billing.

Almost every go-to-market question has a public answer. Who is hiring for the role that implies our product. Which companies are advertising right now. What people complain about when they describe the problem in their own words. What competitors promise, and what their customers say afterwards.

That information is scattered across a dozen platforms, each with a different shape, a different rate limit, and a different reason to make bulk access difficult. So most teams either pay for a database that went stale months ago, or they assign the gathering to a person who deserves better work.

The alternative is usually worse: a set of homegrown scrapers that one engineer understands, which break silently every time a platform ships a redesign.

VoxScrape puts 17 sources behind one interface with one credit balance. You describe what you want, the run executes, and structured results come back — as records you can export, not as a dashboard you have to interpret.

The billing model was the first design decision, not the last. Credits are consumed per result returned. A run that finds nothing costs nothing. A run that fails is refunded automatically. A run cannot exceed the hold you authorised when you started it. Those properties are enforced in the billing code itself rather than promised in a support policy, which is the only version of that promise worth making.

The same reasoning drives the AI layer. Scoring, enrichment, and message drafting are metered per row rather than bundled into a seat price, so running them on ten records costs roughly a tenth of running them on a hundred, and you are never paying a monthly fee for a feature you tried once.

There are three layers. Collection gathers publicly accessible records from a chosen source — a search, a set of profile URLs, a set of business listings. Enrichment runs optional AI passes over those records to score them against your ideal customer profile, fill in missing fields, detect buying intent, or draft an opening message. Simulation goes further, modelling how a described audience is likely to react to a message before you send it.

Every layer is optional and separately metered. Plenty of customers use only the first, treat the export as the product, and never touch the AI features. That is a perfectly good way to use it.

Public means public. VoxScrape collects information that is already visible to anyone browsing the web. It does not log into accounts on your behalf, defeat access controls, or retrieve anything from behind a paywall.

Your exports are yours. We do not sell customer data, we do not aggregate exports into a resale database, and we do not use the contents of your runs to train models. The business is funded by subscriptions, which is the arrangement least likely to end with your data as the product.

Tools do not excuse conduct. Collecting public data lawfully is straightforward; contacting people afterwards is where the obligations live. The acceptable use section is explicit about what the Service may not be used for, and we would rather lose a customer than be the tool behind a spam operation.

VoxScrape is designed, built, and operated by Damien Vernon, through infin8 LLC, a company registered in the United States. There is no support team standing between you and the person who wrote the code — when you email support, that is who replies.

Being small shapes the product. Features ship when they are genuinely useful rather than to fill a comparison table, sources are removed when they stop returning quality results, and the roadmap is driven mostly by what customers report as broken or missing.

If you have found a bug, want a source added, or think something here is wrong, get in touch — or write directly to support@voxscrape.com.