Honest AI cost reporting
Where Vigil draws the line between a number it can stand behind and one it cannot, and why. Our dashboard shows a dash, not $0.00, wherever a cost is unknown: a zero is a claim and a missing figure is an absence, so each dash carries its reason, and every rate carries a provenance label.
The same rule decides what we say about ourselves. A proxy in the request path records what crossed the wire, an SDK reports what your code believes happened, and each knows things the other cannot. These posts set out the rule and the cases that tested it.
Start here
LLM cost accuracy: why we show a dash, not $0.00
The rule itself, case by case.
All 5 posts

Sceptical about a proxy in front of your AI? Good. Here's how to check us.
A proxy sees your AI traffic in transit. Here is exactly what Vigil does with that position, what it never does, and seven ways to verify it yourself.

Vigil and GDPR, in plain language
Who is responsible for what, where Vigil's data lives, which companies help us run it, and the rights you have under GDPR. No legal fog.

What Vigil stores, and what it never stores
Vigil never stores your prompts or the model's responses. The full list of what it records per AI call, where it lives, and how long it is kept.

Proxy vs SDK observability: in-band and out-of-band
Proxy vs SDK observability: an SDK reports what your code believes happened, a proxy records what crossed the wire, and each knows what the other can't.

LLM cost accuracy: why we show a dash, not $0.00
A zero is a claim and a missing figure is an absence. For LLM cost accuracy, our dashboard shows a dash wherever a plausible number would have been a lie.
The same rule runs the public token cost calculator: no verified rate, a dash.