Pricing
Priced on usage, not guesswork
Each plan bundles a monthly credit allowance, then bills extra usage per credit — so light users aren't subsidizing heavy ones, and pricier models (Claude, ChatGPT) don't quietly lose us money. Placeholder numbers for this demo; swap in your real costs before selling it.
Demo — Free
This site — try the product with no commitment.
Starter
For a single site testing real usage.
Includes 500 credits/mo, then $0.06/credit
≈ 500 messages/mo on Gemini (Balanced)
- Knowledge base
- 8 pages / ~300 KB
- History retention
- 14 days
- Bots / sites
- 1
- RAG (your content)
- —
- ✓No RAG — your whole knowledge base is sent with every message, so it's kept small on purpose
- ✓Your own knowledge base (not the shared demo one)
- ✓Email support
Growth
For a site that gets real support volume.
Includes 3,000 credits/mo, then $0.03/credit
≈ 3,000 messages/mo on Gemini (Balanced)
- Knowledge base
- 150 pages / 10 MB
- History retention
- 30 days
- Bots / sites
- 3
- RAG (your content)
- Included
- ✓RAG grounded in your own docs/site content — a big knowledge base doesn't cost more per message
- ✓Conversation history & handoff to a human
Scale
For high-volume or multi-site deployments.
Includes 10,000 credits/mo, then $0.02/credit
≈ 10,000 messages/mo on Gemini (Balanced)
- Knowledge base
- 750 pages / 50 MB
- History retention
- 90 days
- Bots / sites
- 10
- RAG (your content)
- Included
- ✓Priority model access (no shared queueing)
- ✓Up to 10 bots on one account
Enterprise
Committed volume, custom limits, and an SLA.
Committed volume pricing
≈ — messages/mo on Gemini (Balanced)
- Knowledge base
- Custom
- History retention
- Custom
- Bots / sites
- Custom
- RAG (your content)
- Included
- ✓Custom integrations & SSO
- ✓Dedicated support & SLA
- ✓On-prem / VPC options
Estimate your cost & find your plan
Pricing is usage-based because different models cost different amounts to run — tell us roughly what you need and we'll estimate the monthly bill and point you at the right plan.
Do you need your own knowledge base?
Pricing FAQs
Why isn't this a flat "unlimited messages" price?
Every message costs us real money — an AI model call, and (on paid plans) a retrieval lookup over your knowledge base. A flat unlimited price either gets abused by a handful of high-volume accounts and loses money, or has to be priced so high that light users overpay for it. Usage-based pricing means you pay for what you actually use, and it means we stay in business.
What's a "credit"? Why not just count messages?
A Claude or ChatGPT reply costs us roughly 8x what a Gemini reply does, and Groq is cheaper still. Charging every plan the same price per raw message would make Premium-model usage lose money. Credits fix that: Balanced (Gemini) = 1 credit/message, Economy (Groq) = 0.5, Premium (Claude/ChatGPT) = 8. Your plan's included allowance and overage rate are both in credits, so the price always tracks what a conversation actually cost.
What happens if I go over my included credits?
Nothing breaks — overage is simply billed per credit at the rate shown for your plan. That overage rate runs higher than your plan's bundled rate on purpose: it protects us against usage spikes we didn't plan capacity for, and if you're consistently paying overage, upgrading to the next tier down will almost always be cheaper than staying put. You can also set a monthly spend cap so it stops (or alerts you) instead of billing past a limit you choose.
Why does Starter have a smaller knowledge base than Growth?
Starter doesn't use retrieval (RAG) — it sends your entire knowledge base as context on every single message, which is simple but means a bigger knowledge base directly means a bigger bill. We cap it small on purpose to keep Starter predictable. Growth and up switch to RAG, which only pulls in the relevant snippets per question, so a much larger knowledge base barely moves the cost per message.
Can I switch plans later?
Yes, upgrade or downgrade at any time — your embed script tag doesn't change, only the limits and rates behind it.