Tokens at cost: what is in a Thaliq statement
Some platforms resell you the model with a margin on every token. A Thaliq paid plan is a monthly fee plus the tokens you use, at what they cost us, with no margin for now.
There is a quiet tax in some AI agent platforms: they resell you the model. You send a message, it hits their account, and they bill you the tokens plus a margin. On a small pilot it is invisible. At production volume it is a line item that grows with your success.
Thaliq bills it differently. A paid plan is a monthly fee for the platform, plus the tokens you use at what they cost us. For now, we add no margin on those tokens.
What is in a statement
Everyone starts on the Free plan, which has no fee and includes its tokens, within its limits. When you move to a paid plan, the console closes a statement at the end of each month with:
- the plan fee;
- a line per model, with the tokens it used (input, output and prompt cache) at the model’s price;
- any credits we granted you, and the taxes that apply to your organization.
You read each statement in the console and download it as a CSV file. We invoice you from it: there is no automatic card payment. To move to a paid plan, write to [email protected] and we change your plan for you.
What “at cost” means
Tokens are billed at the model provider’s list price, and everything a provider charges for your usage counts: your agents’ turns, the embeddings of your knowledge bases and the runs of your Evals. The hosted service calls the model with Thaliq’s own accounts, so there is no key for you to manage or bring. Today it runs Claude Haiku 4.5 on Amazon Bedrock.
Why give up the easy money?
A margin on tokens would make us earn more when you spend more on the model, which is the wrong incentive for an infrastructure company. We would rather be paid for the platform being worth paying for: the loop, the tools, durable runs, approvals, Evals, Lens and the channels. The statement shows the tokens apart from the plan fee, so you can always see which is which.
Your agents and your data stay yours
Your agents are YAML manifests that you keep in your own Git repository and apply with the CLI, so they are yours to move. The owner of an organization can export its data from the console, conversations, usage and statements included, and ask for it to be deleted. And Thaliq can run in your own infrastructure, where an agent’s file names the provider and the model: changing them is an edit to spec.model.
Jhosetp Chino
We write about how we build Thaliq: the decisions, what went wrong and what we changed. From Lima.