Home/about
About

What BlockTracer is, and what it costs you

A block explorer whose transaction pages are debugging sessions. Arriving at a transaction with a published trace means arriving in its execution — not at a button that opens one.

The read path

The browser reads static files and nothing else. No API, no database, no third-party endpoint, no per-request compute of any kind. A page load is a series of GETs against an object store behind a CDN, and most of those objects are immutable and therefore free after the first visitor in a region.

Mutable objects
One per chain — the generation pointer. Everything else a page reads is written once and never modified, which is why there is no coherence protocol and no cache to get wrong.
Chains published
1
Third parties
None
Account
Not required to read anything

The trust strip, itemised

No account
Every published trace opens anonymously. An account is needed only to ASK the pipeline to generate a trace that does not exist yet, because that costs us compute — and the result is then public for everyone.
No ads
There is no advertising surface, and no market-data widget either: each would cost a scheduled request per visitor and none helps someone who arrived with a transaction to understand.
No tracking
No third-party requests at all, so there is nothing to track you with. What this deployment can observe is its own CDN logs.
Complete history
The pipeline builds the address index, so every listed chain has complete history and complete log coverage. There is no capability to negotiate and no degraded variant of an address page.
No record caps
History is stored as immutable segments keyed by block range and paged by walking them, so the thousandth page costs what the first one does. Nothing is truncated at a row count.

What this product will not do

Guess
Where a fact is not published, the page says so and says what would make it appear. An unconditional claim about variable values is the one thing this product cannot afford.
Fall back to a node
No page falls back to a live tracing endpoint, and no page computes anything itself. Both would reintroduce a third-party dependency and a second source of truth.

What this deployment can observe →