Sign in

Generation,
uncomplicated

A convenient, cost-effective service for generating video, images and music from a single balance.
Top AI models at low prices. Works without a VPN.

Get startedWhat the platform can do
model catalog
  • Seedance 2.5new
  • Gemini Omni 1.1new
  • Kling
  • Nano Banana
  • GPT Image
  • Grok Imagine
  • Claude

One balance for every direction

Video, images and sound are made in one interface and charged from one account. No separate provider subscriptions or keys needed.

Video

A clip from a text description, animating a still image, stitching a first and last frame. References are added with @.

12 models

Images

Key frames, product shots, covers. You can edit a finished image with words instead of rebuilding it from scratch.

10 models

Music

A track or background bed for a clip: genre, mood and instruments are set in words.

3 models

Upscale

Raise the resolution of a finished video or image before publishing.

2 models

Chat with AI: one model or several at once

Claude, Gemini, DeepSeek and Grok in one window. Ask one model or line up to four and compare answers to the same question.

One question

Come up with a name for a workout-diary app and explain your choice

Claude

"Approach". A gym word that's also about how a person approaches things: short, distinctive, no translation needed.

Continue in this chat
Gemini

"Traction". Sport plus growth: sounds tech-forward, easy to find in the store, not taken in the local market.

Continue in this chat
DeepSeek

"Rack". The barbell sits on the rack, and so does your progress. A soft, everyday word instead of aggressive fitness jargon.

Continue in this chat
Two to four models side by side
The prompt goes to every column at once. Answers are typed out in parallel, each in its own column.
Any answer can be continued
A column opens into a regular conversation, and you can ask a follow-up without leaving the row.
Files in the conversation
Images, PDFs, audio, video and text. What a model accepts is listed on its card.
Pay with tokens
Chat spends tokens, not credits. The approximate message price is shown before you send it.

Let the agent write the prompt

Describe your idea in one line. The agent shows what's missing and puts together a detailed prompt for the chosen model.

Your description

a cat barista brewing coffee in a small coffee shop

Seedance 2.0, one shot, 10 seconds

Analysis

34/100

The idea is clear, but the scene isn't described

  • Describe the character: coat color, apron, facial expression
  • Add light and setting: morning, steam, warm lamp
  • Set the camera movement and the scene's sound
Ready prompt

Close-up: a ginger cat barista in a linen apron froths milk behind the counter of a small coffee shop. Morning light from the window, steam and dust motes in the beam, a brass espresso machine on the counter. The camera slowly pushes in from the paws to the face, the cat squints contentedly. Warm film palette, muted jazz and the hiss of steam in the background.

Writes for the model you chose
Seedance, Veo, Kling and Lyria each want something different. The agent knows what you'll generate with and shapes the prompt for it.
Takes references into account
If images or video are attached, appearance and location are already set by them. Then the tips are about character movement, emotion and camera work with sound.
Remembers your settings
One shot or several, duration, aspect ratio, output language. It all makes it into the prompt text.
Works outside the studio too
The same agent is called via the API and the Prompt Agent node in ComfyUI. It doesn't run the generation itself — you only pay for the model that writes the text.

The team works from one account

The owner hands out credits to members, generations pile up in shared projects, and spend is visible by person and project right away, not at month's end.

Team projectLaunch promo

18 395credits spent

left 31 605

  • IIldarOwner8 24064 generations
  • MMarinaAdmin6 11541 generations
  • TTimurMember3 04527 generations
  • ZZhenyaMember9959 generations

These numbers are an example. In the studio, these breakdowns are based on the team's real generations.

  • One account for everyone

    The owner tops up the team and hands out credits to members. Unused credits can be pulled back in one click.

  • Cost per project

    Spend is tracked by project, member and model. You know the cost of the work before you bill the client.

  • Shared generation feed

    Everything the team made, filterable by member and project: status, settings, amount charged.

  • Roles and invitations

    Owner, admin and member. Invitations go by e-mail, no separate accounts or card transfers needed.

API and ComfyUI nodes

The same models, the same balance and the same projects from your own code or straight from a ComfyUI graph.

Start a generationDocumentation
curl https://synthez.app/api/v1/videos \
  -H "Authorization: Bearer sk_live_..." \
  -d '{
    "model": "dreamina-seedance-2-0",
    "prompt": "a cat barista brewing coffee @loft",
    "duration": 10,
    "resolution": "1080p",
    "projectId": "prj_7f2a"
  }'

  • Team key

    Generations land in the team's project and are charged to the member's sub-account. The pipeline runs within budget, not around it.

  • Keys under control

    Multiple keys, an expiry date, instant revocation. Each one shows spend and last use.

What the key can access
  • POST/v1/videosvideo
  • POST/v1/imagesimage
  • POST/v1/musicmusic
  • POST/v1/promptprompt agent
  • POST/v1/estimateprice before starting
  • POST/v1/uploadsupload a reference
  • POST/v1/upscaleupscale up to 4K
  • GET/v1/generations/{id}status and result
  • GET/v1/projectskey's projects
  • GET/v1/balanceaccount balance

Node package for ComfyUI

Video, images, music, upscale and the prompt agent, right in the graph. Our server does the work, no GPU of your own needed, and the credits are the same as in the studio. The list of models and their settings are pulled from the server, so the nodes know exactly what the studio knows. The archive is downloaded from the "API" section: unpack it into custom_nodes and put your key into the "API Key" node. Inside are a README and ready-made workflows.

  • API Key
  • Video
  • Image
  • Music
  • Upscale
  • Prompt Agent
  • Upload File
  • Load Video → Frames
  • Refresh Models
  • Balance

How the studio works

References, run history, projects and a shared team balance are all in one interface. Generations run in parallel and the library stays fast.

  • References with @

    Up to nine images, three videos and three audio files in one generation. The same character or location can be reused in the next shots.

  • Projects and library

    Generations and references are sorted into projects. A frame from a good generation is added to the library in one move and then works as a reference.

  • History and rerun

    Every run is saved with its status and the amount charged. A successful generation can be rerun with the same settings.

  • Parallel generations

    Video and images are processed at the same time: you can start the next generation without waiting for the previous one. The queue doesn't block your work.

  • The library stays fast

    History and references open instantly even with thousands of generations — the feed scrolls without lag or minute-long loading.

  • Price shown before you start

    Cost is calculated from the model, duration and quality, and shown on the run button. You agree to the amount in advance, not after the fact.

Only the models people actually use

The top model in each direction plus a couple of fast ones for drafts. All of them run from one balance, and new versions are added on our side.

  • Seedance 2.5

    BytePlus Seedance

  • Seedance 2.0

    BytePlus Seedance

  • Seedance 2.0 Fast

    BytePlus Seedance

  • Seedance Mini

    BytePlus Seedance

  • Kling 3.0 Omni

    Kling AI

  • Kling 3.0

    Kling AI

  • Kling 3.0 Turbo

    Kling AI

  • Gemini Omni 1.1 Flash

    Google Gemini

  • Veo 3.1

    Google Veo

  • Veo 3.1 Fast

    Google Veo

  • Grok Imagine 1.5

    xAI Grok

  • Grok Imagine

    xAI Grok

  • Nano Banana 2

    Gemini 3.1 Flash Image

  • Nano Banana Pro

    Gemini 3 Pro Image

  • Nano Banana 2 Lite

    Gemini 3.1 Flash Lite Image

  • Seedream 5.0 Pro

    ByteDance Seedream

  • Seedream 5.0 Lite

    ByteDance Seedream

  • GPT Image 2.5 Sunburst

    OpenAI

  • GPT Image 2.5 Flare

    OpenAI

  • GPT Image 2

    OpenAI

  • Grok Imagine 2.0

    xAI Quality

  • Grok Imagine Fast

    xAI Grok

  • Lyria 3

    Google Lyria

  • Lyria 3 Pro

    Google Lyria

  • Seed Audio 1.0

    BytePlus Voice

  • Iris HQ

    Topaz Video

  • Astra 2

    Topaz Video

Pay for generations, not a subscription

The balance is topped up with credit packages for rubles, and a charge happens for each specific run. Packages and the current rate are shown on the top-up page.

What a run's cost is made of

Model
the rate depends on the provider
Duration
seconds of video or track
Quality
frame resolution or render mode
References
input images and videos can affect it too

The amount is calculated before you start and shown on the run button. Every charge is saved in the generation history.

Answering the main questions

Try it on your own task

Sign up, top up your balance and run your first generation. No subscription and no minimum term.

Get started