No plans and no subscription. Start with 50,000 free credits ($5), valid for 90 days, then pay per task from a prepaid balance of credits. Failed calls are refunded.
Worth $5 and valid for 90 days. Then pay per task at 1 credit = $0.0001. The API lists its own credit costs, so you always see the current ones.
For teams that need BeEmotion models on their own devices, or a deployment planned with us.
List the current credit cost of every task from the API at any time.
curl https://api.beemotion.app/v1/tasks \
-H "Authorization: Bearer $BEEMOTION_API_KEY"Credits
Each organisation has a prepaid balance of credits, and 1 credit = $0.0001. Calls draw it down, and there is no subscription.
Image calls
An image call costs the sum of the credits of the tasks it asks for, whatever the number of faces or detections. Expression, gaze and face recognition add face detection automatically, counted once per call. An image call for expression costs 5 + 5 = 10 credits.
Video jobs
A video costs the frames analysed (duration × frames per second, rounded up) times the sum of the per-frame credits of its tasks. One minute of face detection and expression at 1 fps costs 120 credits; 2 fps costs twice that.
Chat
Chat is charged per token and per image, times the model multiplier, rounded up to a whole credit. Each response reports its tokens, credits and cost_usd.
Storage and data transfer
From 1 November 2026, stored data and downloads are charged in credits on the 2nd of each month for the previous month. Uploads are free. Each organisation gets the first 1 GB-month of storage and the first 1 GB of downloads free every month.
Refunds
If a call fails, its credits are refunded to the balance automatically. Unused balance is not refundable, except where the law requires it.
Top-ups and bonus
Organisation admins top up by card, from $25 to $1,000, through Stripe Checkout. Larger top-ups add credit: $250 buys $262.50 of credit, and $1,000 buys $1,100. Paid credit never expires.
Free credit
New accounts get 50,000 free credits ($5) once. It is used before paid credit and expires 90 days after it is granted.
Credits per image, and per analysed video frame. A call costs the sum of the tasks it asks for, whatever the number of faces or detections. 1 credit = $0.0001.
| Task | Per image | Per video frame |
|---|---|---|
| Face detection | 5 | 1 |
| Expression | 5 | 1 |
| Gaze | 5 | 1 |
| Body pose | 5 | 1 |
| Object detection | 5 | 1 |
| Face recognition | 5 | 1 |
| Captions | 3 | – |
| OCR | 3 | – |
| Open-vocabulary detection | 6 | – |
| Open object detection | 4 | – |
| Identity enrolment | 5 | – |
| Video transcript | – | 1 |
Expression, gaze and face recognition add face detection automatically, counted once per call. An image call for expression costs 5 + 5 = 10 credits. A dash means the task does not run on that kind of input.
From 1 November 2026, the data your organisation stores and the data it downloads are charged in credits. Uploads are free.
| What | Credits | Cost |
|---|---|---|
| Storage | 250 credits per GB-month | $0.025 per GB-month |
| Downloads from origin | 1,050 credits per GB | $0.105 per GB |
| Downloads from the CDN cache | 350 credits per GB | $0.035 per GB |
| Uploads | Free | Free |
1 credit = $0.0001. The example bills below include a month of storage and downloads.
Credits = text input tokens ÷ 1,000 × 1, plus output tokens ÷ 1,000 × 20, plus images × 2, times the model multiplier, rounded up to a whole credit.
| Rate | Credits |
|---|---|
| Text input | 1 credit per 1,000 tokens |
| Output | 20 credits per 1,000 tokens |
| Image | 2 credits per image |
| Model | Multiplier |
|---|---|
| Vision chat model | 1.00 |
| Example | Text tokens in | Tokens out | Images | Credits | Cost |
|---|---|---|---|---|---|
| Short question50 words in, 2 sentences back, no image | 80 | 60 | 0 | 2 | $0.0002 |
| Long answera paragraph in, a 400-word answer back | 250 | 512 | 0 | 11 | $0.0011 |
| Three images3 photos and a question, a paragraph back | 60 | 200 | 3 | 11 | $0.0011 |
The chat guide has the reservation rules and streaming details. Chat pricing
Worked examples at the current credit costs.
| Example | Credits | Cost | Bought in a $1,000 top-up |
|---|---|---|---|
| Retail / venue video10 cameras, 12 h/day, 30 days, 1 fps, face + expression + gaze | 38,880,000 | $3,888.00 per month | $3,534.55 |
| Ad testing200 participants × 2-minute webcam clip, 5 fps, face + expression + gaze | 360,000 | $36.00 per study | $32.73 |
| App, 10,000 expression calls10,000 image calls, face + expression | 100,000 | $10.00 per month | $9.09 |
| App, 100,000 expression calls100,000 image calls, face + expression | 1,000,000 | $100.00 per month | $90.91 |
| App, 1,000,000 expression calls1,000,000 image calls, face + expression | 10,000,000 | $1,000.00 per month | $909.09 |
| Multimodal10,000 captions + 5,000 OCR + 20,000 chat requests (200 tokens in, 150 out) | 125,000 | $12.50 per month | $11.36 |
| Storage and downloads10 GB stored and 20 GB downloaded from origin in a month, from 1 November 2026 | 22,200 | $2.22 per month | $2.02 |
"Bought in a $1,000 top-up" is what the same credits cost when bought that way: a $1,000 top-up buys $1,100 of credit.
Prices below were read from each provider's public price list on the date shown in each row, and each row links the price page. Providers change their prices, so check the source before you rely on a figure. All amounts are US dollars at list price.
| What | BeEmotion | Provider price | Source |
|---|---|---|---|
| AWS Rekognition DetectFacesFace detection, which can return face attributes. US East (N. Virginia). We counted face detection + expression. | $0.0010 per imageSame price | $0.0010 per image | AWS Rekognition pricingRead |
| Azure AI FaceFace detection, Standard tier, first million transactions. East US. We counted face detection. | $0.0005 per imageAbout 2 times lower | $0.0010 per image | Azure AI Face pricingRead |
Provider prices are for the first million images a month.
| What | BeEmotion | Provider price | Source |
|---|---|---|---|
| AWS Rekognition SearchFacesByImageSearches a face in the image against stored faces, first million images. US East (N. Virginia). We counted face detection + face recognition. | $0.0010 per imageSame price | $0.0010 per image | AWS Rekognition pricingRead |
Our price includes the face detection that recognition needs.
| What | BeEmotion | Provider price | Source |
|---|---|---|---|
| AWS Rekognition stored video, face detectionFace detection on stored video. US East (N. Virginia). We counted face detection. | $0.006 per video minuteAbout 17 times lower | $0.100 per video minute | AWS Rekognition pricingRead |
| AWS Rekognition stored video, face searchFace search on stored video. US East (N. Virginia). We counted face detection + face recognition. | $0.012 per video minuteAbout 8 times lower | $0.100 per video minute | AWS Rekognition pricingRead |
We bill each analysed frame, so a minute at 1 frame per second is 60 frames. Provider prices are for stored video.
| What | BeEmotion | Provider price | Source |
|---|---|---|---|
| OpenAI gpt-4o-miniStandard tier, input tokens. We counted Vision chat model, text tokens. | $0.10 per 1 million input tokens$0.05 lower | $0.15 per 1 million input tokens | OpenAI API pricingRead |
| Google Gemini 2.5 Flash-LiteStandard paid tier, text, image and video input. We counted Vision chat model, text tokens. | $0.10 per 1 million input tokensSame price | $0.10 per 1 million input tokens | Gemini API pricingRead |
| Google Gemini 2.5 FlashStandard paid tier, text, image and video input. We counted Vision chat model, text tokens. | $0.10 per 1 million input tokensAbout 3 times lower | $0.30 per 1 million input tokens | Gemini API pricingRead |
Text tokens only. We bill each image as a flat charge on top, listed in the chat rates above.
| What | BeEmotion | Provider price | Source |
|---|---|---|---|
| OpenAI gpt-4o-miniStandard tier, output tokens. We counted Vision chat model, text tokens. | $2.00 per 1 million output tokensAbout 3 times higher | $0.60 per 1 million output tokens | OpenAI API pricingRead |
| Google Gemini 2.5 Flash-LiteStandard paid tier, output including thinking. We counted Vision chat model, text tokens. | $2.00 per 1 million output tokensAbout 5 times higher | $0.40 per 1 million output tokens | Gemini API pricingRead |
| Google Gemini 2.5 FlashStandard paid tier, output including thinking. We counted Vision chat model, text tokens. | $2.00 per 1 million output tokens$0.50 lower | $2.50 per 1 million output tokens | Gemini API pricingRead |
Our output price is higher than two of the three listed and lower than the third.
| What | BeEmotion | Provider price | Source |
|---|---|---|---|
| AWS S3 StandardStorage, first 50 TB a month. US East (N. Virginia). We counted storage, charged on the month’s average. | $0.0250 per GB-month$0.0020 higher | $0.0230 per GB-month | Amazon S3 pricingRead |
| Azure Blob Storage Hot LRSStorage, first 50 TB a month. West Europe. We counted storage, charged on the month’s average. | $0.0250 per GB-month$0.0054 higher | $0.0196 per GB-month | Azure Blob Storage pricingRead |
Our storage and download prices apply from 1 November 2026. Our first GB-month each month is free for every organisation. Provider prices are for the first 50 TB a month.
| What | BeEmotion | Provider price | Source |
|---|---|---|---|
| AWS Data transfer out to the internetFirst 10 TB a month after the free 100 GB. US East (N. Virginia). We counted downloads served from origin. | $0.105 per GB$0.015 higher | $0.090 per GB | Amazon S3 pricingRead |
| Azure Internet egressMicrosoft global network routing, first 10 TB a month after the free 100 GB. West Europe. We counted downloads served from origin. | $0.105 per GB$0.018 higher | $0.087 per GB | Azure bandwidth pricingRead |
Our first GB of downloads each month is free, and downloads served from our CDN cache cost less, as listed above. AWS and Azure each give the first 100 GB a month free.
Each row compares like with like: where a provider price includes more or less than ours, the row says what we counted. Prices for a provider and region not listed here may differ.
1 credit = $0.0001, so $25 buys 250,000 credits. GET /v1/tasks lists the current credit cost of every task. Billing and credits guide
The 50,000 sign-up credits expire 90 days after they are granted. Free credit is used before paid credit. Paid credit never expires.
Calls are refused with a 402 insufficient_balance error until an organisation admin tops up the balance. Your data is kept: the monthly storage and transfer charge can take the balance below zero, and data is never deleted because of a negative balance.
Storage is the data your organisation keeps in BeEmotion Cloud, charged on the average amount stored over the month. Data transfer is the data you download from it. Downloads served from the CDN cache cost less than downloads from origin, and uploads are free. Each organisation gets the first 1 GB-month of storage and the first 1 GB of downloads free every month. These charges start on 1 November 2026.
No, except where the law requires it. Failed calls are always refunded.
Chat is charged by what it uses: input tokens, output tokens and images. Longer messages, longer answers and more images cost more, and each response reports what it used. Chat pricing
Every response includes usage with the units processed, the credits charged and cost_usd.
BeEmotion models also run on edge devices. Get in touch to talk about an on-device or private deployment. Contact us
The live demo runs the models on your own webcam, with no account needed.