Token pricing for each tool

Approximate token costs for video, audio, photo translation, captions and chat image models.

Written By Umakhan Magomedov

Last updated 3 days ago

This page lists the approximate token cost for each AI tool in VocaLingo. The app always shows an estimate before you start, so you can verify the cost for your specific input.

ℹ️ All prices are estimates. Actual charges may vary slightly based on final processing results.

Translate Video

Mode	Price
HeyGen (lip-sync)	About 3.67 tokens/sec
HeyGen (audio only)	About 3.67 tokens/sec
HeyGen + Enhanced Cloning	About 7.34 tokens/sec
ElevenLabs dubbing	About 3.67 tokens/sec
ElevenLabs dubbing + Lip Sync	About 4.34 tokens/sec

Add Captions

Mode	Approximate cost
Mirage animated templates	15 tokens per started minute
Standard subtitles	Speech recognition + rendering
Translated subtitles	Speech recognition + translation + rendering

Learn more: How Add Captions works.

Translate Audio

Charged per pipeline step. Voiceover is optional and runs when you tap play on the Translation tab. Full settings guide: Translate Audio settings.

Speech recognition

Provider	Price
ElevenLabs Scribe (default)	About 0.0133 tokens/sec
OpenAI Transcribe	About 0.02 tokens/sec
Whisper	About 0.01 tokens/sec

Example: 60-second voice message with ElevenLabs Scribe ≈ 0.8 tokens.

Translation

Model	Price
Gemini Flash Lite	0.028 tokens/1K chars
Gemini 3.8 Flash (default, first pipeline translation)	0.056 tokens/1K chars
GPT 6 Luna	0.008 tokens/1K chars

First upload always uses Gemini 3.8 Flash. Settings model applies to re-translations only.

Voiceover (TTS)

Provider	Price	Extra fees
ElevenLabs	0.01 tokens/sec	None
OpenAI	0.03 tokens/sec	None
MiniMax	0.15 tokens/sec	150 tokens first clone per voice
Qwen	0.15 tokens/sec	Minimum 5 tokens per request
HeyGen v3	About 3.67 tokens/sec	None

Example: 30-second voiced translation with ElevenLabs ≈ 0.3 tokens. Same text with HeyGen ≈ 110 tokens.

Photo translation

Model	Typical cost
Nano Banana 2	About 8 tokens per photo
GPT Image 2.5 Sunburst	About 6 tokens per photo
GPT Image 2.5 Flare	About 6 tokens per photo

Used in Text Translator photo mode. Chat image generation uses the same family of models. Details: How to translate text from a photo.

Other tools

Speech to Text, Video to Text, Text Analysis, Text to Speech, Tourist Translator, Text Translator, AI Calls and AI Chat show a token estimate before you start. Cost depends on duration, selected model, text length and optional summaries.

Frequently asked questions

Where can I see the exact cost before I submit?

Each tool shows a token estimate before you tap the action button. In Translate Audio, the estimate updates when you change settings or generate voiceover.

Why is Translate Audio cheaper with ElevenLabs than HeyGen?

ElevenLabs uses standard synthetic voices (0.01 tokens/sec). HeyGen clones the speaker voice with higher quality processing (about 3.67 tokens/sec) and takes longer to generate.

Do I pay for voiceover if I only read the translation?

No. Voiceover tokens are charged only when you tap play to generate audio. Recognition, translation and summary from the automatic pipeline are charged when processing completes.

Why was I charged 150 tokens for MiniMax?

The 150-token fee is a one-time clone activation per voice. It applies the first time MiniMax clones a new voice from your file or saved Custom Voice. Later generations with the same voice cost only 0.15 tokens/sec.

Was this helpful?

Still need help? Share an idea