Synthflow uses usage-based billing, meaning you only pay for what you use. This page explains how we measure usage, how we price it per plan, and how charges appear on your invoice.
Phone calls are billed per second, starting when the human and the AI agent connect and ending when the call terminates. Partial seconds are rounded up to the nearest second.
If a transfer completes successfully, billing stops at the point of transfer. Usage after the transfer is not billed.
Rather than rounding each individual call, we aggregate seconds across calls. Each whole minute that accumulates is reported as a separate billing event on your invoice, so you’re charged fairly for your actual usage. All invoices are available in the Billing Portal under Settings > Plan & Billing.
Here’s how it works:
In this example, four calls totaling 125 seconds produce 2 billing events on your invoice (one per accumulated minute), with the remaining 5 seconds carried over to the next call.
Calls with a no-answer or hangup_on_voicemail status are billed at a flat rate of 5 seconds at your plan’s per-minute rate. Calls with a failed status are not billed.
Every call in Synthflow has one of the following states:
Calls marked as paused, spam, or in-progress are not charged on your invoice; they only become billable when they reach a terminal state.
Chat is converted into billable minutes: every 5 AI-generated messages = 1 billable minute, charged at your plan’s per-minute rate.
On Pay As You Go, calls are billed per LLM at the following rates:
Separately from LLMs, the voice engine is billed on every PAYG phone call. The voice engine is the combination of your transcriber (speech-to-text) and synthesizer (text-to-speech) providers. The Synthflow Voice Engine is billed at $0.09/min.
Legacy plans use a single per-minute rate that combines LLM and voice engine usage. Billing is not broken down by LLM; you only see billable minutes used.
Enterprise customers have custom rates based on their agreement.
Plans that include a monthly minute allowance (e.g. 6,000 minutes) only charge for usage above that allowance. If your total minutes used in the period stay within the included amount, that usage is not billed; only overage is charged. This is clearly reflected on your invoices.
When rollover is enabled, any unused included minutes at the end of a contractual period roll over to the following period instead of expiring. This allows high-volume accounts to absorb fluctuations in usage without losing the value of their included allowance.
For example, if you finish the current period with 50,000 unused minutes, those minutes carry over and are added to your included allowance at renewal, giving you more to draw from in the next period.
Rollover minutes is an Enterprise feature available on demand. It is not included in every Enterprise plan. Contact your account manager to enable it for your account.
For Pay As You Go and legacy plan customers, we use billing thresholds to ensure timely invoicing as your usage grows. Enterprise plans use custom billing arrangements instead.
Billing thresholds are triggers, not spending caps. When your unbilled usage reaches $200, an invoice is automatically generated and charged to your payment method. The invoice includes all usage accrued up to the moment it is created, which will always be equal to or greater than $200.
Thresholds are not evaluated in real time. Instead:
Because of this, the invoiced amount can exceed the threshold, especially for high-throughput accounts.
A $200 threshold does not mean your invoice will be exactly $200. If usage grows quickly between threshold checks, the invoice will reflect the full amount accrued at the time of creation. This is expected behavior. If you need hard spending limits, you must implement usage caps on your own. Billing thresholds will not enforce a ceiling.
A customer with a $200 threshold accrues usage rapidly. The system evaluates the threshold at 10:00 AM and finds usage has already reached $458. An invoice is generated for $458, the full amount accrued (not $200).
After the invoice is created, usage tracking resets and continues accumulating toward the next threshold or month-end. You may receive multiple invoices in a single billing period if your usage is high.
Your current minute usage is displayed in the sidebar, giving you visibility into your consumption at a glance.
The Usage dashboard is available under Settings > Usage. It shows:
Usage shown reflects billable credits only. Credits used during call transfers are not included and may cause differences from other analytics views.


For subaccounts, call and chat usage is measured the same way as for the parent workspace (minutes, aggregation, and the Usage dashboard). The per-minute rate that applies to each subaccount is determined by the pricing plan that subaccount is subscribed to on your Stripe connected account.
Calls with a failed status are not billed. Calls with a no-answer or hangup_on_voicemail status are billed at a flat rate of 5 seconds at your plan’s per-minute rate, instead of full per-second billing.
Your current minute usage is shown in the sidebar. For detailed usage history (including total events, total cost for a period, and breakdown by date and by provider), use the Usage dashboard under Settings > Plan & Billing. Invoices and full billing history are in the Billing Portal there.
Usage on the dashboard reflects billable credits only. Credits used during call transfers are not included, which can cause differences from other analytics views.
If your usage exceeded the $200 billing threshold, an additional invoice was generated mid-cycle. This is normal for high-volume usage and helps avoid large end-of-month charges.
Billing thresholds are triggers, not caps. Thresholds are checked periodically (roughly every 30 minutes), so usage can grow beyond $200 before an invoice is created. The invoice covers all usage accrued up to that point. This is expected behavior, not an overcharge.
Usage alerts are not yet available but are on our roadmap. For now, monitor your usage through the sidebar indicator.
International calls may have different per-minute rates depending on the destination. Check your specific rates in Settings > Plan & Billing or contact support for details.
Billing starts when the human and the AI agent connect and stops when the call ends or a transfer completes successfully. Milliseconds are rounded up to the nearest second.