AI support that
doesn't fight you.

No per-resolution billing. No hard caps. No enterprise sales calls. Just a chatbot that answers your customers.

yoursite.com

Welcome to Acme

How can we help you today?

ShippingReturnsPricing
Where's my order #4821?
Order #4821 shipped Tuesday via UPS. Tracking says it arrives Friday. Want me to email you the link?
Yes please
Sent ✓ Anything else I can help with?

Why most AI support is a bad deal.

Three patterns that make every existing tool feel hostile.

01

Per-resolution billing

The better your bot works, the more you pay. A good month turns into a $2,000 invoice.

02

Hard-capped "free" tiers

Free until you actually get traction. Then a 100-message wall, and an upgrade prompt.

03

Sales calls for pricing

No price on the page. Book a call, sit through demos, get quoted $30k/year minimum.

Up and running in five minutes.

1

Paste your docs

A URL, a PDF, or pasted text. The bot learns your product, pricing, and FAQs.

2

Match your brand

Colors, tone, name. Pick the model that fits — we run Llama, Mistral, Qwen, DeepSeek.

3

Drop one line of code

A single <script> tag on your site. The widget is live.

Flat monthly. No surprises.

Cancel anytime. Self-hosted open-source models. Zero data retention.

Starter
$9/mo

Side projects and small stores.

  • 1 site
  • 1,000 chats/mo
  • Self-hosted OSS model
  • Embed widget
Join waitlist
Growth
$49/mo

Past 100 tickets a day.

  • 1 site
  • 25,000 chats/mo
  • Custom prompt tuning
  • Slack/Discord alerts
  • 40+ languages
Join waitlist
Scale
$99/mo

Multi-site, high volume, teams.

  • 3 sites
  • 100,000 chats/mo
  • Team seats (5)
  • API access
  • Dedicated support
Join waitlist

Join the waitlist.

First 500 teams get 100 AI chats free every month, forever. Then 20% off for life when we launch. No card to sign up.

🔒 Zero data retention. We run open-source models on our own GPUs. We never store your docs or your conversations.

No card. No spam. We'll only email you when we launch + once with your invite.

FAQ

Do you really never store my data?

Yes. Your docs are embedded in-memory, used to answer, then discarded. Conversation transcripts are not logged. There's no analytics database with your customers' questions in it. We run open-source models on our own GPUs — no third-party LLM API ever sees your text.

What model do you use?

Open-source models — Llama, Mistral, Qwen, DeepSeek — hosted on our own infrastructure. You can pick the default per bot, and Pro+ can mix models for different intents.

How are you able to offer flat pricing?

Self-hosted models on our own GPUs means our marginal cost is hardware and electricity, not per-token API fees. We pass that through as flat monthly pricing, with a soft cap. If you consistently exceed your plan's chats, we show you a banner and offer a transparent upgrade — no surprise invoices.

What's the catch on "Join the waitlist"?

No catch. One email when we launch, and the first 500 teams get 100 chats free every month for the lifetime of their account. After that, 20% off for life.

Can I cancel anytime?

Yes. One click, no email loop, no retention call. Because we don't store your data, there's nothing to export — it was never persisted in the first place.