HugoCurio.

What Curio costs

Usage-based, priced from our real token cost, and shown before you spend a cent. No seats, no monthly floor, no per-domain fee. Here's exactly what a unit is and what it costs.

The unit is a conversation

One visitor, one chat thread, is one conversation. Every question and follow-up inside it belongs to that same conversation. You're billed for the model tokens each answer actually uses, and nothing for the questions Hugo never gets asked.

During your free trial a whole conversation — every follow-up included — is $0.

What one conversation costs

Standard
~$0.09
per conversation · live average
billed at 5.2× our token cost
Pro
~$0.14
per conversation · live average
billed at 8.5× our token cost

These are the live running average from our own usage, not a list price. Your actual cost moves with how long a conversation runs. The same number drives the counter on the homepage.

The three ways you pay

Free trial

$0
30 days, no card
  • 5 conversations a day, every follow-up free
  • The full portal, nothing walled off
  • 10-page crawl of your site
  • Install, test, and go live before paying a cent
Most sites start here

Standard

~$0.09 / conversation
5.2× token cost · prepaid credits
  • Grounded answers, each one cited to its source
  • Live human handoff and lead capture
  • Daily refresh that re-reads only what changed
  • Unlimited domains and seats
  • 10-page crawl per site

Pro

~$0.14 / conversation
8.5× token cost · prepaid credits
  • Everything in Standard, plus:
  • 200-page crawl per site
  • Transcript email on every chat
  • Collector / webhook forwarding

How the meter works

We pay Anthropic for the model and Voyage for embeddings. You pay that model cost times your plan's multiple, drawn from a prepaid balance. Here are the raw per-million-token rates we're billed, so you can check the arithmetic yourself:

ModelInput / M tokensOutput / M tokens
Claude Haiku 4.5$1.00$5.00
Claude Sonnet 4.6$3.00$15.00

Multiply the tokens an answer uses by these rates, then by 5.2× (Standard) or 8.5× (Pro). That is the whole meter. When your prepaid balance reaches $0, Hugo pauses politely and tells the visitor — it never runs up an overage or a surprise invoice.

What the multiple pays for

The gap between our raw token cost and your price isn't margin alone. It covers the embeddings and the daily re-crawl that keep answers current; the free trial nobody pays for; the credit we hand back automatically when we have downtime; the help desk and live human handoff; and the hosting, with enough left over to keep building. We would rather show you the multiple than bury it in a "contact sales" quote.

Pricing you can check is a promise, not a courtesy. If any number on this page is stale, or the math doesn't add up, tell Hugo right here. Holding ourselves to a price you can verify is the whole point — it's written into our constitution.

Start free →Compare with other tools

Per-conversation figures are the live running average from our own usage and vary with conversation length. Token rates are set by the model providers and change when they do; your exact usage and prepaid balance are always in your portal. The 30-day trial needs no credit card.