Chinese AI video, image and text models. Big ideas. One API key.
Call Seedance, Kling, Wan, Vidu and DeepSeek models from one FishToken account. Pay as you go from a prepaid balance, and keep your OpenAI SDK code: change the base URL and the API key.
Video / Image / Text
Featured models
Four video model families and one text model to start with. Each model page lists the model IDs and an example request.
Seedance 2.0 Fast
VideoVideo generation from the Seedance 2.0 line. The standard Seedance 2.0 and Seedance 2.5 model IDs are available too.
Model IDsdoubao-seedance-2-0-fast-260128doubao-seedance-2-0-260128doubao-seedance-2-5-260628
/v1/video/generationsModel page Kling V3
VideoVideo generation with two model IDs: the standard kling-v3 and the kling-v3-omni variant.
Model IDskling-v3kling-v3-omni
/v1/video/generationsModel page Wan 2.7
Video + ImageText-to-video, image-to-video, reference-to-video and video editing, plus two image generation models.
Model IDswan2.7-t2vwan2.7-i2vwan2.7-r2vwan2.7-videoeditwan2.7-imagewan2.7-image-pro
/v1/video + /v1/imagesModel page Vidu Q3
VideoVideo generation with three model IDs: viduq3, viduq3-turbo and viduq3-mix.
Model IDsviduq3viduq3-turboviduq3-mix
/v1/video/generationsModel page Text and code through the chat completions endpoint: captions, product copy, support replies.
Model IDsdeepseek-v4-flash
/v1/chat/completionsModel page New here?
Make a first call
Create a key and send a first request in Python, Node.js or curl. The quickstart also lists common errors.
QuickstartWhy FishToken
One account instead of many
Working with several Chinese video models usually means several sign-ups, several top-ups and several SDKs. FishToken puts them behind one key.
Several video models, one account
Seedance, Kling, Wan and Vidu can all be called with the same FishToken API key. No separate registration or top-up with each model developer.
One balance, pay per use
Top up once and spend the same balance on any supported model. You are charged per call, the balance does not expire, and redemption codes are supported.
Works with the OpenAI SDK
The endpoint is OpenAI-compatible. Point base_url at https://fishtoken.ai/v1, use your FishToken API key, and keep the rest of your code.
Southeast Asia
Where it fits
Common jobs for teams selling and publishing in Indonesia, the Philippines, Vietnam and the rest of Southeast Asia.
Product videos for online shops
Turn a product photo into a short clip with an image-to-video model such as wan2.7-i2v, or describe the shot in a prompt.
Social media content
Generate short vertical clips for short-video and social channels, then edit and caption them in your usual tools.
Ad creative testing
Produce several versions of the same scene, switch models by changing one field, and test which version your audience responds to.
Customer service and copy
Use deepseek-v4-flash to draft support replies, captions and product descriptions. Check output quality in Bahasa Indonesia, Filipino or Vietnamese before it reaches customers.
Quickstart
Your first call in about 30 seconds
Create an account
Sign up at fishtoken.ai/register with a username, a password and an email verification code.
Add balance and create a key
Top up or redeem a code, then create an API key in the console. Store it as
FISHTOKEN_API_KEY.Run the code
This example calls the text model. Node.js and curl versions are in the quickstart.
from openai import OpenAI
import os
client = OpenAI(
base_url="https://fishtoken.ai/v1",
api_key=os.environ["FISHTOKEN_API_KEY"],
)
resp = client.chat.completions.create(
model="deepseek-v4-flash",
messages=[
{"role": "user", "content": "Write a two-line caption for a handmade batik tote bag."}
],
)
print(resp.choices[0].message.content)FAQ
Questions
- What is FishToken?
- FishToken is an API platform. You create one account, get one API key, and call AI models for video, image and text from Chinese model developers through an OpenAI-compatible endpoint.
- Which models can I call?
- This site features Seedance 2.0 Fast, Kling V3, Wan 2.7, Vidu Q3 and DeepSeek V4 Flash. The current model list and prices are on fishtoken.ai/pricing.
- Do I have to change my code?
- If you already use an OpenAI-compatible SDK, change two settings: set the base URL to
https://fishtoken.ai/v1and use your FishToken API key. Video models use the/v1/video/generationsendpoint; each model page has an example. - What do I need to sign up?
- A username, a password and an email address to receive a verification code.
- How does billing work?
- You add balance first, then each call is charged to that balance. The balance does not expire, and you can also add balance with a redemption code. See how pricing works.
- Does FishToken store my prompts?
- FishToken records call metadata such as the model, token counts and time for billing and logs. The content of your requests is not stored by the platform. Your request is forwarded to the upstream model service so it can be processed.
- Are there rate limits?
- Yes: 60 requests per minute for standard accounts and 120 requests per minute for accounts that have topped up.
- What content is not allowed?
- No sexual content, no graphic violence, no real person's likeness without consent, and nothing illegal or infringing. The full list is in the content rules.
- How do I get help?
- Email support@fishtoken.ai.
Get your API key
Create an account, add balance, and make your first call from the quickstart.