Per seat. Never per resolution.
Every plan runs the full console. Shinobi's managed engine is metered in plain AI actions with a real allowance, or bring your own Anthropic, OpenAI, Gemini or Grok key on any plan and pay us nothing for AI.
Try it, build on it, stay small
- 1 seat, the full core console
- 50 AI actions a month on Shinobi's engine
- Bring your own model key, even here
- Widget shows "Powered by Shinobi"
Small shops, one to three people
- Widget, inbox, knowledge brain, cases, customers
- Shopify + WooCommerce, workflows, tool cards
- Reports and QA on every conversation
- 200 AI actions per seat per month, pooled
The default for growing teams
- Everything in Starter
- Sensei, Training, Scout, Self-Healing fixes, Missions, Experiments, Strategies, Versions, Publish
- SMS + WhatsApp (your Twilio), Slack bridge, CRM handoff, send-as email
- 500 AI actions per seat per month, pooled
Brands and wholesalers, still self-serve
- Everything in Growth, unlimited seats
- 2,000 AI actions per seat per month, pooled
- Priority support and migration help
- SSO, audit logs and dedicated capacity as they ship
One AI action = one customer conversation Shinobi handles on our engine (all turns, web, SMS or WhatsApp), or one background job you ask for: a Training sim, a Scout answer, a drafted reply, a brain build. Automatic quality riders (QA review, case classification, embeddings) never count. Allowances pool across the workspace; past them Shinobi keeps answering and overage bills at $0.15 per action. Alerts land at 80% and 100% of the allowance, the Usage page shows the live meter, and the raw ledger exports as CSV. Bring your own key and the AI line is $0, on every plan.
Usage pricing exists only where we carry a real cost: storage past your plan's allowance is $0.20/GB-month, and outbound email on Shinobi's sender past the allowance is $1 per 1,000 (connect your own Gmail or Outlook and it is free). SMS and WhatsApp run through your own Twilio account, so carrier costs are yours at cost, never marked up. That is the whole list: three meters, no quota maze.
Add-ons
Capabilities that sit on top of your plan, sold as apps in the App Store. Ours ride the same rails third-party builders will. Prices are announced now, buttons appear when each one ships. No fake toggles, ever.
Guided tours in the widget: Shinobi points at parts of your page, narrates, and waits while the visitor drives. You build and publish every tour; tours never click or type for anyone.
The built-in CRM layer: your customer book with accounts and orders, Cases with full activity, and Team Chat with the Slack bridge and huddles. Nothing to configure. It is native.
Your AI partner as a full workspace: chat and cowork over your live business, notebooks, global search, scheduled jobs, and 50+ console operations. Every step traced, everything behind your review gates. Included on Growth and up.
Build forms for your website in the Studio builder. Pick where each form's submissions land (PrintflyOne Cases work best; HubSpot and Salesforce too) and map every field to the record it fills.
Why we never charge per resolution
A per-resolution tax prices your bot like a slot machine: the vendor celebrating every resolved ticket on your invoice is telling you they are surprised it worked. We price seats because Shinobi resolving things is the baseline, not the jackpot. You should celebrate your containment rate, not budget-fear it. We still measure resolution and containment everywhere (Reports, Experiments, Self-Healing); we just refuse to bill on a number a vendor gets to define.
| PLATFORM | MODEL | AI PRICE | THE CATCH |
|---|---|---|---|
| Shinobi | Seats + a plain AI-action meter | $0 on your key | None. The meter is readable, exportable, and never gates the bot. |
| Intercom Fin | Seats + $0.99 per "outcome" | $0.99/outcome | Outcomes include handoffs and disqualifications, not just resolutions |
| Zendesk AI | Seats + per automated resolution | $1.50–$2.00/res. | An LLM verification step decides what you owe |
| Gorgias | Ticket bundles + AI per resolution | ~$1.26 effective | An AI resolution also consumes a helpdesk ticket |
| Sierra / Decagon | Outcome contracts, enterprise only | reported $95K+/yr | No pricing page, implementation fees, annual lock-in |
Questions
Nothing visible to your customers. Shinobi never stops answering. You get an email at 80% and at 100% of the allowance, the Usage page shows the meter, and actions past it bill at $0.15 each on your next invoice. Adding a seat raises the pool immediately.
Yes: Anthropic, OpenAI, Gemini or Grok, on every plan including Free. Keys are stored envelope-encrypted per workspace and validated with a live test call, you pay your provider directly, and your AI line with us is $0. You can even mix: run chat on your key and route background jobs to ours, per surface.
14 days of Growth-level features on every new workspace, no card required. When it ends you land on Free. Everything you built stays saved, and gated features pick up exactly where they left off when you upgrade.
Yes. The price is on this page and checkout is a button, no sales call required. Talk to us only if you need custom work: migrations, custom components, special integrations. That is a services conversation, not a hidden price list.
Signup to a live branded widget takes about 15 minutes: verify your email, paste a key or use ours, brand the widget, add one script tag, or install the Shopify app and skip the tag entirely.