Data sources and access

Where Finitizer Tokenomics gets its numbers, how each dollar is labeled, who can see what, and what Tokenomics cannot see.

Everything in Tokenomics is computed from one set of rows, read from sources you already have.

Where the data comes from#

Source	What it covers	What you set up
Cloud bills Finitizer already reads	Gemini and Vertex AI from the Google Cloud billing export, and Amazon Bedrock from AWS	Nothing new
Nightly pulls	Claude Enterprise, the Claude Console, OpenAI, GitHub Copilot, Cursor, OpenRouter, and an Azure Cost Management export	An administrator stores a key. For Azure, a read-only link to the export's storage container
Gateway logs	Cloudflare AI Gateway request logs. They split each provider's bill by user, team or key, and add no spend of their own	Connect the gateway
Exports you import	22 file formats from providers, clouds and developer tools. The list is below	Upload the file

A stored key is sealed on the server and never shown again.

The import formats include:

Every dollar carries its basis, so you always know how firm a number is:

Actual: read from an invoice.
Allocated: a share of an invoice, split by a rule or a gateway log.
Estimated: a list-price estimate, used where a feed reports tokens but no dollars.

Role	What they can do
Viewer	Read every screen
Engineer and above	Make changes
Finance, admin and super admin	Book value claims, and see spend and activity by person
Administrator	Manage provider keys

Everyone else sees the same totals with people counted, not named.

Tokenomics says so on the page rather than reading a gap as zero:

A channel with no data, a model with no price, or a cost month not yet read is shown as missing.
Edge and desktop models send nothing to read.
A forecast cannot see a launch or a contract that nobody entered.

It also never acts on your behalf:

Finitizer Tokenomics reads AI spend from the bills and exports you already have. Get in touch to see it on your own data.