Priced the Way AI Actually Costs

Every AI product here is a base fee plus a metered component. That is deliberate. Speech recognition, model inference, and speech synthesis all cost money per interaction, so a flat monthly fee would mean heavy users are sold at a loss and light users overpay. Metering keeps the entry price low and the pricing honest at both ends. And because AI Agent Builder is bring-your-own-key, you can hold the model contract yourself and cut our metered rate out of the inference entirely.

Website Chat Widget

$15 per seat / month
Minimum 3 seats

Embeddable, brandable chat widget for your site, routing to an AI agent or through to WhatsApp.

About Website Chat Widget

AI Agent (chat)

$99 per month
Plus a per-message rate agreed with our sales team

LLM-powered agents across WhatsApp, SMS, RCS, Telegram, and web, with your own model key. The per-message rate is set case by case against your model choice, message complexity, and expected volume.

About AI Agent Builder

Unified Omnichannel Inbox

$129 per month
Any 3 channels included, plus $19 per agent seat

One inbox across every channel, so your team answers from a single place rather than app by app. Additional channels $29 each per month.

Voice AI

$199 per month
Plus a per-minute rate agreed with our sales team

Conversational voice agents. Metered per minute because speech recognition, inference, and synthesis all carry real cost. The per-minute rate is set case by case against your languages, destinations, and expected call volume.

Workflow Automation Builder

$199 per month
Any 3 channels included. Add-on at Growth, included at Scale and Enterprise

Build automated journeys and triggers across channels. Includes the customer journey builder. Additional channels $29 each per month.

Agentic AI / Custom Development

$5000 project, plus retainer
Scoped per engagement

Custom agent development against your own systems, workflows, and data, with an ongoing support retainer.

Prices in USD, exclusive of applicable taxes. All figures are starting points, and final pricing depends on your message volume, channels, and whether you bring your own model key.

Pick Any 3 Channels. Add More for $29.

The Unified Omnichannel Inbox and the Workflow Automation Builder each include any three channels of your choice at the listed price. You decide which three, so a retailer running WhatsApp, Instagram, and Facebook Messenger pays exactly the same as an enterprise running WhatsApp, RCS, and Telegram.

WhatsApp
RCS
Telegram
Instagram
Facebook Messenger
More being added

How additional channels are charged

Beyond your included three, each additional channel is $29 per month. If you run both the Inbox and the Workflow Builder, a shared additional channel is charged once, not twice. As new channels are added to the platform they become selectable on the same terms, and we will not reprice your existing selection because the list grew.

Bring your own model key, and pay your provider directly

AI Agent Builder connects to your own OpenAI, Anthropic, Azure OpenAI, or Google Gemini account. Your key is encrypted at rest, you hold the model contract, and you decide which provider ever processes a customer query. When you bring your own key, inference is billed to you by your provider rather than metered by us, which matters both for cost control and for any compliance review that asks where customer data is processed. See data residency for why that question comes up.

What Sits Where

Product Pricing model Availability
Website Chat Widget Per agent seat, minimum 3 seats Any platform plan
AI Agent (chat) Base plus a per-message rate, agreed case by case with sales Growth upward
Unified Omnichannel Inbox Platform fee plus per agent seat. Any 3 channels included, $29/mo per additional channel Growth upward
Voice AI Base plus a per-minute rate, agreed case by case with sales Growth upward
Workflow Automation + Journey Builder Flat add-on. Any 3 channels included, $29/mo per additional channel Add-on at Growth, included at Scale and Enterprise
Agentic AI / custom development Project fee plus retainer Scoped per engagement

Messaging channels used by these products are billed on the channel volume ladder, separately from the AI fees above.

Tell us your AI message volume

Rough numbers are fine. We will turn them into a real monthly figure rather than a range, and tell you whether bringing your own model key works out cheaper.

Get a Quote

FAQ

AI Pricing Questions

Because our cost is genuinely variable. Speech recognition, model inference, and speech synthesis all cost money every time the agent replies. A flat monthly fee would mean heavy users are sold at a loss and light users overpay and leave. Base plus metered usage is the only structure that stays fair to both, and it is the reason we can keep the entry price low.

Each message the AI agent generates and sends to the end user. Inbound messages from the customer are not billed, and neither are replies typed by a human agent after an escalation. We meter per message rather than per conversation because a long, complex conversation costs us more to run than a short one, and per-message billing reflects that honestly in both directions: simple exchanges cost you less. The exact counting basis is confirmed in your agreement so it matches what your reports show.

Yes, and most clients do. AI Agent Builder is bring-your-own-key for OpenAI, Anthropic, Azure OpenAI, and Google Gemini. Your key is encrypted at rest, you hold the model contract, and you control which provider ever processes a customer query. If you bring your own key, model inference is billed by your provider rather than by us.

Any three channels of your choice are included with each of them, from WhatsApp, RCS, Telegram, Instagram, and Facebook Messenger. Additional channels are $29 each per month. You pick which three, so a business running WhatsApp, Instagram, and Facebook Messenger pays the same as one running WhatsApp, RCS, and Telegram.

No. Each product comes with its own three included channels, but where the same additional channel is used by both, it is charged once rather than twice. Adding Instagram to an account running both products costs $29 a month in total, not $58.

The channel list grows over time. New channels become selectable as part of your included three, or as a paid addition at the same $29 rate, and we will not reprice your existing selection because the list got longer.

For anything that sends on a messaging channel, yes, because the channel runs on your platform deployment. See platform pricing. The website chat widget can run against a web-only deployment.

No. It is bundled with Workflow Automation, because buyers reasonably see them as one capability and splitting them would just create confusion at quoting time.

Because both vary too widely for a single published number to be honest. The AI message rate depends on which model you run, how complex your agent's replies are, and your expected volume. The voice minute rate depends on your languages, the destinations you are calling, your monthly minutes, and the speech and model providers involved. Rather than publish a range wide enough to be useless, we agree the rate with you before you sign and write it into your agreement. Tell our sales team your volumes and we will give you a firm figure.

They are starting points. Final pricing depends on your message volume, the channels involved, and whether you bring your own model key. Tell us roughly how many AI messages a month you expect and we will give you a real figure rather than a range.

See Platform Plans and Capacity Tiers

View Platform Pricing

The Fine Print, Up Front

Pricing Terms

Currency and tax

All prices are in US dollars and exclusive of applicable taxes, duties, and withholding. Indian rupee figures shown against channel volume pricing are indicative and convert at the prevailing rate on your invoice.

Billing and commitment

Subscriptions are billed monthly, quarterly, or annually in advance. Annual prepay includes two months free. Setup is a flat $100 one-time charge on every plan, billed with your first invoice, and waived entirely where you prepay quarterly or annually. Downgrades are not automatic, because your dedicated server and database are provisioned for the tier you buy. Talk to us before committing if your volume is likely to fluctuate.

Fair use, daily threshold and overage

We do not suspend, throttle to zero, or terminate your service for exceeding an allowance. Traffic above your daily threshold is queued and released in the next window, and no charge arises from daily queueing alone. Usage above your monthly allowance is billed as an overage charge in arrears on your next invoice, at the rate agreed with you and recorded in your order form. Overage rates are set against your expected volume and destination mix rather than published as a single figure, and are agreed before you sign, so no overage charge ever arrives at a rate you have not already seen. Overage is charged against the monthly allowance only, never the daily threshold.

How usage is counted

Channel volume on the WhatsApp, RCS, Telegram, and OBD Voice ladder is counted on delivered messages, not submitted. A message submitted but not delivered does not count toward your volume and is not billed. Published ladder rates apply to perpetual licence holders; on a rented subscription plan, channel rates are quoted separately and recorded in your order form. For other usage, the counting basis is stated in your order form and matches what your platform reports show. Our platform records are the primary evidence of usage, save in the case of manifest error.

Notifications and sustained overage

We alert you at 80% and 100% of your monthly allowance, so an overage charge is never the first you hear of it. If you exceed your monthly allowance in three consecutive months, we may ask you to move to the next plan from your following renewal, on thirty days' notice. Any overage accrued in your final month remains payable on termination.

Service levels

Where your plan carries an uptime SLA, it covers availability of the SMPPCenter application, measured monthly. It excludes scheduled maintenance notified at least 48 hours in advance, unavailability attributable to the underlying hosting provider, outages or throttling by any upstream carrier or vendor, failures of your own infrastructure, network, or credentials, and events outside our reasonable control. Where we provide hosting, the infrastructure service level is the hosting provider's own published SLA, named in your order form and passed through without enlargement.

Perpetual licences, upgrades and security

A perpetual licence grants a non-exclusive, non-transferable right to use the version delivered at purchase, in object code form only. No source code rights are granted. The annual upgrade charge (AUC) is mandatory for the first three years and covers version upgrades, feature releases, and security patches. Should AUC lapse thereafter, you retain the right to run the version you own but receive no further upgrades, releases, or security patches, and you would be asked to acknowledge in writing that continued use without security patches is at your own risk. Reinstating AUC is charged in respect of each lapsed year.

Servers and deployment

On a subscription plan we provide and maintain the server infrastructure. On a perpetual licence you provide and maintain it. Where a plan states a number of servers, that refers to servers across which a single deployment of the software is horizontally scaled. It is not a licence to run multiple independent deployments.

Support

Day-to-day support is ticket-based on every plan. Scale and Enterprise include 24/7 support. On every plan, including the entry tier, an emergency telephone line is available where the service is down, and we will restore it regardless of your support tier.

Price changes

Prices and inclusions on these pages may change. Where a change affects an active subscription, we will give you at least thirty days' written notice before it takes effect, and your current pricing holds until your next renewal date. Newer products, including AI and voice, are at an early stage of pricing and are the most likely to be revised as usage patterns become clear.

NCPR, DLT and regulatory data

Every plan includes the scrubbing and validation engine. The underlying NCPR, DND, and DLT data is supplied by you, obtained and maintained under your own registration with the relevant registry or operator, and loaded into the platform by you. SMPPCenter does not source, licence, or warrant the accuracy or currency of that data, and is not responsible for sends permitted or blocked as a result of data you have supplied. Real-time scrubbing against an external provider's API is supported as a chargeable addition. Responsibility for regulatory compliance of your traffic, and of your customers' traffic, remains yours.

Channel and third-party charges

WhatsApp conversation charges levied by Meta, RCS carrier charges, operator and vendor fees, and model inference billed by your own AI provider are separate from our platform fees and passed through. We will break them out separately in your quote.

This is a plain-language summary provided so you know what you are buying before you talk to us. It is not the contract. The binding terms are those in your signed agreement and order form. See also our Terms and Conditions and Privacy Policy. Prices and inclusions on this page are current at the time of publication and may change.

Client Feedback

What Our Clients Say

Our clients run regulated, high-volume messaging infrastructure and prefer not to be named publicly. These testimonials are shared with permission while keeping commercially sensitive details confidential.

High throughput, intelligent routing, and detailed monitoring have significantly simplified our messaging operations.
Operations Manager, Telecommunications Provider Kenya 8 million SMS per month Runs SMPP Gateway Client since 2023
We needed complete control over our messaging infrastructure instead of relying on a hosted provider. The platform has delivered excellent performance and reliability.
Solutions Architect, Software Company South Korea 99.99% service availability Runs Enterprise Messaging Platform Client since 2020
The delivery reports and routing controls give us the visibility we were looking for. The system has integrated well with our internal applications.
Head of Technology, Enterprise Communications Company South Africa 4 operator routes Runs SMPP Server Client since 2021
The migration was completed smoothly and the platform has remained stable under heavy production traffic. The flexibility of self-hosting was one of the biggest advantages for us.
Head of Infrastructure, Financial Services Company India 18 million messages per month Runs SMPP Server Client since 2021

Names and company details are withheld at our clients' request. This is common among telecom providers, SMS aggregators, fintech companies, and enterprises that consider their messaging infrastructure commercially sensitive.

LET US PRICE YOUR AI DEPLOYMENT

Tell us roughly how many AI messages a month you expect and which channels they run on, and we will give you a real figure.

GET A QUOTE