Dial in what you need — panels, conversations, seats, staff, storage — and we'll show the smallest plan that fits + what it'd actually cost you each month.
Prefer a quick chat? Try the price predictor.
Public chatbots, private team assistants, or multi-agent verdict panels. Most businesses start with 1 chat and 0 others.
Chat panels
Public-facing chatbot
Assistant panels
Private team tool
Verdict panels
Multi-agent debate
Flovato Agent can add these for you during setup — this calculator is for power users who want to model cost up front. Don't know yet? 1 chat, 0 assistants, 0 verdicts is a solid starting point.
One conversation = one back-and-forth chat with a customer or your assistant. Drives the token estimate below.
A typical short support chat is ~1.5K tokens; long multi-turn conversations run 5–10K. Run out of plan tokens and overage kicks in (or switch to PAYG).
A seat is anything that logs in: people, kiosks, shared devices, or system accounts. Different from your staff roster — staff are operational records and don't need to log in. Each extra seat above your plan's included count also bumps your monthly token allowance.
Staff roster — people on your team that the system tracks (line cooks, technicians, contractors). Doesn't include people that need a dashboard login (those are seats). Extras run $1–$3/mo each depending on plan — way cheaper than seats.
Text documents — PDFs, FAQs, manuals, vendor catalogs — are AI-searchable, so your chatbot can quote your actual files. Photos and images are stored but not indexed for AI search. Your plan's storage covers all your data — these files plus every record managed across every manager — under one allowance.
AI-searchable storage for your docs, PDFs, and images. When a customer asks “what's your warranty?”, the AI pulls the answer straight from your files.
Not ready to pay?
Try our 14-day trial — card on file, no charge
Smaller caps, but it's genuinely free. The plan below is what you'd pay monthly if you picked it directly instead.
Monthly cost
For small businesses getting started with AI
Numbers assume Haiku-class models (baseline billing). Premium models (Sonnet 5×, Opus 20×) burn proportionally more tokens per conversation — pick a model you like, your meter drains at that rate.
Start with StarterAll plans at this load
Click a plan to lock it and see its breakdown above.
Estimates only. Actual usage varies with conversation length, token choice, and model. Token rollover and overage rates are factored in.