TinyTrack, first public slice
milestone ·
TinyTrack is a sidecar I use to check whether a small web product actually sends the conversion events it declared. I sign into an evidence console, hand it a public URL, and a worker visits that page in a real browser. What comes back is a receipt of the network traffic.
The 21 August 2026 plan is why this exists. It aimed at JS and web SaaS teams of one to ten people, the kind that ship GA4 and paid acquisition pixels with no analytics engineer on staff. The one completed job in that plan was narrow. Prove that a named event left the page when a real browser finished a journey. TinyTrack stays beside the product. If TinyTrack is down, the product still runs. TinyScale is the platform. TinyTrack is the only public story I am telling from it.
I run this on my own work. It is not a launch, and it is not a product strangers can sign up for.
The first result that felt honest landed on 23 August 2026. Representative owned fixture cases passed the browser run. Direct generic event was an expected pass and came back HTTP 204. Consent denied came back inconclusive, with no matching requests. I wanted the quiet result as much as the pass. A page that should stay silent and still phones home is a worse failure than a miss.
On 24 August 2026 I watched a Cloudflare Web Analytics beacon leave a page with HTTP 204. The request happened. That is all I will claim from that run.
A network response proves transport, not attribution, and not destination ingestion. Here is the everyday version. You watch a checkout page. You asked whether a purchase event left for a collector path. The receipt shows HTTP 204. That means the browser sent the hit and got an empty success. It does not mean Meta or GA4 stored it, billed it, or attributed it. I have not proven they stored it.