Pricing

On the Auto model, about $70 to $90 a month for about 3,000 messages. Voice notes, images and PDFs push it toward the top of that range. The cost calculator above gives a live number for your volume and model.

No. $29/month includes $29 of usage and unlocks premium features (custom domain, white-label, audit logs, sub-organizations, Slack support, larger knowledge storage, SSO). Usage above $29 is billed at the same per-message rates as Pay As You Go, you are never charged a premium for going over, you just pay for what you use.

Real money. You pay Invent in dollars. Internally we convert your spend to a "credit" unit so that messages, knowledge searches, and broadcasts can draw from the same balance, but the dashboard, the calculator, and your invoices all show dollar amounts. You never need to think in credits.

Messages are priced per-message based on the AI model you choose. Every model shows its estimated cost per message right next to its name in the assistant settings. That estimate already assumes an average conversation with prompt caching.

Simple text-only messages cost less than messages with voice notes, images, or PDFs, and longer responses cost more than short ones. Your full Instructions and any triggered actions (Knowledge Base Search, integrations) also add to each message.

You can track your spend in real time under Settings → Usage in your dashboard.

The per-message estimate assumes prompt caching is active. Your first message has no cache to read from, so it processes the full context at standard input rates. After that, consecutive messages in the same conversation benefit from caching and typically cost less than the estimate.

Yes. Set a Monthly Spending Limit under Settings → Billing, anywhere from $0 to $100,000. When the limit is reached, billable operations pause until you raise the limit or the next billing cycle begins. You can also enable auto-recharge to keep your assistant running without manual recharges.

Start with Auto, our recommended default. If you need to tune costs down or quality up later, the assistant settings show every available model with its per-message cost next to it, so you can compare at a glance.

Yes. You can change the model at any time in your assistant settings. The new model's pricing applies from that point forward. Each model shows its estimated per-message cost right next to its name.

Yes, prompt caching is enabled on every supported model. When your assistant handles consecutive messages in the same conversation, cached context is reused instead of reprocessed, so those messages cost less than the per-message estimate. Caches last up to one hour.

Free Pay As You Go messages reset monthly based on your organization's creation date. If your organization was created on March 10, your free messages reset on the 10th of each month. Unused free messages do not roll over.

No. Free messages are provided to one organization per account. If you manage multiple organizations, only one will receive the free monthly allocation.

Very little. Indexing is a one-time charge of about $0.0005 per chunk (a chunk is ~350-400 words, about half a page), so a short FAQ page is about $0.001 and a full 100-page help center is about $0.07. Each search your assistant runs at query time is about $0.0005, and identical searches are cached free for 2 hours across your whole organization. For most businesses, knowledge costs are a few cents to a few dollars a month.

Yes. Pay As You Go includes 100 MB of knowledge storage. The Business plan raises this to 2 GB. Contact us for higher limits on the Enterprise plan.

Broadcasts are charged for each message sent, from your credit balance. Email is $0.001 per message, and that is the final price. SMS is $0.001 per message, plus your carrier's delivery cost. WhatsApp is $0.004 per message, plus Meta's template pricing, which Meta bills to you directly.

The number of contacts depends on your plan. Pay As You Go includes up to 500 contacts, and the Business plan supports up to 100,000. Contact us for higher limits on the Enterprise plan.

Yes, you can embed the chat widget on your website with a simple JavaScript snippet. We have step-by-step install guides for Next.js, WordPress, Shopify, Webflow, and more.

Yes. Use the HTTP Request action to connect your assistant to any external API or database. This lets it fetch real-time information during a conversation and answer based on live data instead of just its knowledge base.

Yes. Beyond the system Admin, Developer, Manager, and Agent roles, you can create custom roles with granular permissions, deciding exactly who can manage billing, edit assistants, view analytics, or only reply to conversations.