Most AI support tools charge per seat or per resolved ticket, so your bill grows exactly when your support load does. ansvery runs on your own OpenAI or Gemini key: the AI cost goes to your provider at cost, and we charge nothing for the software.
Paste an OpenAI or Gemini key and run your AI support department for free. You pay us only when you want live agents in the loop, or want us to set everything up and run it for you.
€0
with your own OpenAI or Gemini key
from €199
per month. Custom pricing once live agents are included
Per-resolution pricing hides the unit economics. Per-token pricing does not. Put your own numbers in and see the difference at your volume.
Defaults are an example only, in the range of a small modern model. Check the live rates at OpenAI or Google and put your own numbers in — this is your bill, not ours.
Your estimated AI spend
$2.88/ month
About $0.002 per AI-handled ticket, billed by your provider.
What ansvery charges
$0
No subscription and no ticket limit on the bring-your-own-key tier. You pay us only if you want live human agents or a managed setup.
Compare: per-resolution pricing
Same 1,200 AI-handled tickets would cost $1188 per month. That is $1185 a month more than paying your AI provider directly.
The honest answer is that AI inference is the expensive part of an AI support product, and it is not our cost to mark up. When you bring your own key, that cost sits where it belongs — between you and OpenAI or Google, at their published rate, with no margin added by us. What is left is software, and we would rather have you using it than paying a subscription to evaluate it.
We make money when you want people, not software: live human agents working alongside the AI, or our team configuring and running the setup for you. That starts at €199 per month and is quoted individually, because staffing a support function is not a self-serve purchase. There is no upsell trigger hidden in the free tier, no ticket ceiling, and no feature held back to force an upgrade — the free tier includes every AI agent, both operation modes, the knowledge base, analytics and every integration.
Your AI bill depends on three things: how many tickets the AI handles, how much context each answer needs, and which model you choose. A small modern model answering a support question with knowledge-base context typically uses a few thousand input tokens and a few hundred output tokens per reply, which lands the cost per ticket in fractions of a cent to a few cents. The estimator above lets you set all three, because the only numbers worth trusting are your own.
You control the spend directly at your provider: set a monthly budget cap, switch to a cheaper model, or revoke the key outright. None of that requires our permission, which is the part that per-seat and per-resolution contracts cannot offer.
Still deciding? The FAQ covers how onboarding works and what happens if your key runs out, and the terms spell out what each side is responsible for.