Billing

Billing is plan-based with metered top-ups. Each /agent/ask call debits cost_units from your current period's quota; the value depends on the mode you ask in.

The unit of billing

Every successful /agent/ask response carries a usage block:

{
  "usage": {
    "prompt_tokens": 124,
    "completion_tokens": 256,
    "cost_units": 2
  }
}

prompt_tokens and completion_tokens come from the underlying model provider - they're informational. cost_units is what we charge.

Mode → cost mapping

Mode	`cost_units`	Required plan	What it's for
`fast`	1	Starter+	Quick responses with standard reasoning and core market knowledge. Good for asset checks, narrative pulses, and high-volume financial Q&A.
`smart` (default)	2	Starter+	Stronger reasoning over complex financial questions, with deeper market context. Good for technical analysis, trade setups, and fundamental stock research.
`max`	8	Pro+	Neurobro's most capable tier: deep multi-asset investigations, fundamental analysis with broad market coverage, and the highest-stakes research.

Calling max from a Starter key returns 403 with code: "mode_not_in_plan". Upgrade to Pro to unlock it.

Plans

The full per-plan breakdown - monthly quota, rate limit, key cap, and support tier - lives on the Pricing page. Two rules matter for billing:

An active subscription is required for billable endpoints.
/health is free and works as long as the key itself is valid.

What `402` means

HTTP/1.1 402 Payment Required
Content-Type: application/json

{ "detail": "..." }

You see 402 when:

the account has no active subscription, or
the current period's quota is exhausted, or
a payment failure paused the subscription.

Resolution: open the dashboard → Billing and either subscribe, top up, or fix the payment method. Successful payment unlocks calls immediately.

Programmatic visibility

The dashboard's Usage page summarises billable counters per period. For per-request visibility, read usage.cost_units on each response - that field is the source of truth.