← All posts

Guides

Synthesia API Pricing: What Does a Video Budget Include?

Budget Synthesia API work by the applicable meter. Separate presenter footage, dubbing and generated clips before comparing enterprise video quotes.

To budget Synthesia API usage, identify the API, eligible plan and billable activity before converting a credit balance into video minutes. Finished-video duration is not a universal billing unit. When comparing a quote with Token360 model access, first separate presenter footage, dubbing and generated scene assets so both offers cover the same work.

Enterprise buyers may need personalized onboarding videos, localized training, or generated footage for a product launch. Those jobs can all involve an API, yet they do not purchase the same output. This article uses Synthesia’s published API and billing descriptions to clarify the purchase, then explains where a Token360 model quote can be compared fairly.

Is there one Synthesia API price for every video task?

No. Start with the API and activity your application needs, then confirm the plan that permits it. Generating a finished presenter video, localizing existing footage and rendering a live interactive avatar are different jobs. A subscription headline or an undifferentiated request for “API access” cannot specify the full purchase.

Synthesia’s API introduction distinguishes the Video API, which produces finished files asynchronously, from Interactive Avatars, which streams a live avatar. It also lists Localization, Upload and Billing APIs. Its API access guidance identifies Pro and Enterprise eligibility for the Legacy v2 API. Sources checked November 20, 2026; verify the particular API and account before ordering.

For example, a workflow that fills an approved presenter template from customer records should be specified differently from an application that generates short atmospheric clips from prompts. Neither description is a complete quote yet, but each identifies a concrete task the supplier can confirm.

Keep three questions separate: can the account call the endpoint, what activity consumes credits, and what deliverable the response provides? A key that authenticates successfully answers only part of the first question. The general enterprise video API quote guide covers access and contract scope; this article focuses on Synthesia’s specific billing distinctions.

Does one finished minute equal one billable minute?

Not necessarily. A finished minute may contain presenter footage, reused assets, new scene generation and audio work. Each component can have a different metering rule. Before estimating cost, ask whether the quoted minute refers to finished output, newly generated avatar footage, localized footage or a live conversation, and confirm which endpoint it covers.

The current Synthesia pricing FAQ describes video-creation credits in terms of generated avatar footage rather than final running time. It separately describes dubbing and generated B-roll. It also says edits are charged for new footage when regenerated. Source checked November 20, 2026. Do not automatically apply that general video-creation explanation to every API or an older enterprise agreement.

This matters when two videos have the same length but different production work. A two-minute presentation with an avatar throughout is a different specification from two minutes largely made of existing screenshots. Reusing an approved scene is also different from generating a replacement. Record those differences rather than forcing every item into one “minutes produced” column.

Require a short billing explanation alongside a representative output. Ask the supplier to identify the operation that was charged and the measured quantity. The goal is to map the billing record to the work performed; it is not to infer an exact future bill from one example.

Which published credit rates can you use as a reference?

Use a rate only after confirming its product and account applicability. Synthesia’s Enterprise credit guide publishes different rates for API usage and dubbing, while explicitly warning that the new schedule may not apply to every Enterprise plan. Treat the figures below as a dated reference for discussion, not your contracted price.

Selected Synthesia Enterprise credit references · November 20, 2026
Published activityCredits per minuteConfirm before budgeting
API usage160Applicable API and billable duration
Dubbing with lip sync80Eligible operation and measured minutes
Dubbing without lip sync40Required output and account terms

Source: Synthesia’s Enterprise credits guide. The account-specific warning is part of the source, not an optional qualification.

For illustration, if your agreement confirms the 160-credit rate and a workload contains 100 billable API minutes, that component requires 16,000 credits. This is hypothetical arithmetic, not 100 guaranteed finished minutes or a dollar quote. Obtain the agreed cash price and eligible credit allocation before converting that requirement into a purchase amount.

A dollar-per-credit estimate obtained by dividing a subscription price by its balance is an allocation assumption, not necessarily a marginal rate. The subscription may fund other usage or capabilities. Explain what has been allocated to this workload and avoid counting an included allowance again as a separate charge.

How should you estimate revisions and language versions?

Build the estimate from the production steps your team will actually request. Separate the first version, language versions, new generated assets and later changes. Then map each step to the supplier-confirmed meter. The approved final runtime alone does not describe how much work was generated, repeated, reused or localized.

Illustrative purchase worksheet — fill with your measured quantities
WorkstreamRecord separatelyQuestion for the supplier
Initial presenter videosTemplates, scripts and required presenter segmentsWhich generated duration is billed?
Language versionsTarget languages and translation/dubbing methodIs this localization, dubbing or a new generation?
Supporting scene assetsModel, duration, resolution and required audioAre these charged separately, and in what units?
RevisionsChanged sections and the operation run againWhat is reused, and what is newly billable?
DeliveryDownload, embedding, hosting and destination requirementsWhat is included after generation?

Suppose a team publishes a short onboarding video in several languages and updates a product name later. Do not assume the change means either a full new charge or no charge. Identify whether it modifies a caption, narration, avatar segment or generated scene, then ask which operation the application will run. This is an illustrative planning scenario, not a customer result.

Testing should capture both the delivered file and the usage change. For a production integration, include the path from job submission to completed asset retrieval. Synthesia’s API introduction documents asynchronous generation and polling or webhook completion; those steps still need an application owner. API availability does not remove that integration work.

The corporate training workflow guide covers content approval and lesson quality. Keep those acceptance requirements in the project brief, while this worksheet identifies the separate billable work needed to produce the lesson.

Can you compare Synthesia with Token360 per video minute?

Only after matching the deliverable and the work included. A price for producing a scripted presenter video is not automatically comparable with the price of generating a scene clip. Token360 model access can be evaluated for the components your application needs, but a lower component price alone does not establish a cheaper finished video.

Token360’s video API page describes prompt- or image-based clip generation and asks buyers to check model-specific settings and billing units. Synthesia’s API documentation describes finished scripted avatar files as one of its API outputs. Sources checked November 20, 2026. These are documented workflow distinctions, not a quality or latency benchmark.

Match the requested result before comparing offers
Your required resultRelevant purchasing scopeWhat must not be assumed
Scripted presenter videoPresenter, voice, template and finished-file workflowAny video-generation endpoint supplies this complete result
A generated scene for an existing editorSelected model, input, clip settings and returned assetPresenter creation, assembly and distribution are included
A complete multilingual campaignAll required components, assembly, localization and deliveryOne unit price covers every production stage

Nor should Synthesia be described as “avatars only”: its current pricing materials include model-generated B-roll. If comparing that component, specify the same model version, settings and required output, then confirm the applicable commercial terms. The provider name is not the comparison unit.

A company may also purchase both: an authoring or presenter workflow and separately generated assets for it. Verify upload support, permitted use, accepted formats and the actual handoff before assuming that combination works. For the broader purchasing decision, use AI software versus a model API. This article does not claim Token360 resells Synthesia or provides its avatar system.

What should enterprise procurement approve?

Approve a written workload and commercial scope, not an isolated headline rate. The offer should identify the API, permitted operations, billable units, applicable credit or usage terms and the delivered result. Ask each supplier to price the portion it actually supplies, with remaining production work assigned to an owner.

For Synthesia, request confirmation of your account’s API eligibility, the current rate schedule and the cash commitment that funds your planned workload. Ask which balances and operations are eligible, how exhaustion is handled and which renewal or expiry terms govern the agreement. Leave unresolved terms out of a claimed saving.

For Token360, provide the exact model workload: inputs, output settings, monthly expected generation and peak demand. Token360 uses tiered enterprise quotes—higher usage receives lower quoted rates—and does not charge a separate service fee. That does not mean free generation or that avatar authoring, editing, localization and delivery are included in every model quote.

The Token360 enterprise offering covers text, image, audio and video model requirements with custom commercial scope. Confirm the actual model route and required features with the team before committing. Source checked November 20, 2026. No feature parity or universal price advantage over Synthesia is claimed.

The decision becomes clearer once both offers name the same output. If you need a presenter workflow, evaluate that workflow. If you need model-generated scenes inside a system you already operate, evaluate those scenes and their integration. Price each against an accepted result, with the remaining work visible.

Written by Yuki Yu for Token360. Documentation checked November 20, 2026. No comparative generation test, customer bill or measured saving is presented. Confirm current account-specific terms before purchase.

  • Enterprise AI
  • API

Bring your next idea to life.

Create a video scene in Creator, or explore models for images, audio, and text to find the right capabilities for your project.

Get started