Vidu AI Video Generation via Fhddos Gateway

Vidu is a powerful AI video and audio generation service you can call directly through Fhddos. Use your existing Fhddos token — there is no separate Vidu account registration required. All endpoints share the /vidu prefix, and billing comes straight from your Fhddos account balance.

Authentication

Add your Fhddos token to every request as a Bearer credential:

Authorization: Bearer <fhddos-token>

Base URL: https://aiapi.fhddos.com

Core capabilities

Capability	Description
Text-to-video	Generate a video from a text description
Image-to-video	Animate a still image into a dynamic video
First / last frame	Generate a smooth transition video between two keyframe images
Reference-to-video	Generate video anchored to 1–7 reference images
Video replacement	Replace the subject in an existing video
Lip sync	Add audio or text dubbing to an existing video
E-commerce one-click	Produce an ad video from a product image in one call
Controllable audio effects	Compose multi-segment audio events on a timeline

Available models

Image / First-last frame
Text-to-video
Reference-to-video
Audio effects

Model	Notes
`viduq3-turbo`	Latest Q3 turbo
`viduq3-pro`	Latest Q3 pro
`viduq2-pro-fast`	Q2 pro, accelerated
`viduq2-pro`	Q2 pro
`viduq2-turbo`	Q2 turbo
`viduq1`	Q1 standard
`viduq1-classic`	Q1 classic
`vidu2.0`	Version 2.0 (first-last frame supported)
`vidu1.5`	Image-to-video only — first-last frame not supported

Model
`viduq3-turbo`
`viduq3-pro`
`viduq2`
`viduq1`
`vidu2.0`
`vidu1.5`

Model
`viduq2`
`viduq1`
`vidu2.0`
`vidu1.5`

Model	Use case
`audio1.0`	Timeline-based multi-segment controllable audio

Billing

Fees are deducted from your account balance per task.
Different models, resolutions, and durations carry different prices.
Enable off-peak mode (off_peak=true) for a 50 % discount on generation cost.
Tasks that fail or are blocked by content review are not charged.

Quick start

Text-to-video

curl -X POST "https://aiapi.fhddos.com/vidu/ent/v2/text2video" \
  -H "Authorization: Bearer $TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "viduq3-turbo",
    "prompt": "Cinematic sunset at sea",
    "audio": true,
    "duration": 5,
    "resolution": "720p",
    "style": "general"
  }'

Query task status

curl "https://aiapi.fhddos.com/vidu/ent/v2/tasks/$TASK_ID/creations" \
  -H "Authorization: Bearer $TOKEN"

Video generation endpoints

All generation endpoints follow the same pattern:

POST https://aiapi.fhddos.com/vidu/ent/v2/{action}

Mode	Path
Text-to-video	`/vidu/ent/v2/text2video`
Image-to-video	`/vidu/ent/v2/img2video`
First-last frame	`/vidu/ent/v2/start-end2video`
Reference-to-video	`/vidu/ent/v2/reference2video`

Common parameters

Parameter	Type	Required	Description
`model`	string	✅	Model name (see tables above)
`duration`	int	❌	Duration in seconds; valid range depends on model
`resolution`	string	❌	`540p`, `720p`, or `1080p`
`movement_amplitude`	string	❌	Camera motion: `auto`, `small`, `medium`, `large`
`off_peak`	bool	❌	`true` for 50 % off-peak discount
`watermark`	bool	❌	Embed a watermark in the output
`wm_position`	string	❌	Watermark position
`wm_url`	string	❌	URL of the watermark image
`callback_url`	string	❌	Webhook URL for task-state change notifications

Text-to-video

Generate video purely from a text prompt. Supports general and anime styles, and audio-sync (audio) on Q3 models.

curl -X POST "https://aiapi.fhddos.com/vidu/ent/v2/text2video" \
  -H "Authorization: Bearer $TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "viduq3-turbo",
    "prompt": "A girl runs toward the camera and winks with a smile.",
    "audio": true,
    "duration": 5,
    "resolution": "720p",
    "style": "general",
    "off_peak": false
  }'

Text-to-video exclusive parameters:

Parameter	Required	Description
`prompt`	✅	Text description, up to 2 000 characters
`style`	❌	`general` (default) or `anime`
`audio`	❌	`true` to generate audio alongside the video (Q3 series only)

Image-to-video

Animate a still image. Supply exactly one image URL in the images array.

curl -X POST "https://aiapi.fhddos.com/vidu/ent/v2/img2video" \
  -H "Authorization: Bearer $TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "viduq3-turbo",
    "images": ["https://your.cdn.com/first.png"],
    "prompt": "The astronaut waved and the camera moved up.",
    "audio": true,
    "duration": 5,
    "resolution": "720p",
    "off_peak": true
  }'

Image-to-video exclusive parameters:

Parameter	Required	Description
`images`	✅	Array of image URLs — must contain exactly 1 image

First-last frame

Generate the middle section of a video between a start frame and an end frame. Supply exactly two image URLs — first frame then last frame.

curl -X POST "https://aiapi.fhddos.com/vidu/ent/v2/start-end2video" \
  -H "Authorization: Bearer $TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "viduq2-pro",
    "images": [
      "https://your.cdn.com/start.png",
      "https://your.cdn.com/end.png"
    ],
    "duration": 5,
    "resolution": "1080p"
  }'

First-last frame exclusive parameters:

Parameter	Required	Description
`images`	✅	Array of exactly 2 image URLs: `[startFrame, endFrame]`

Reference-to-video

Generate a video that preserves subjects from 1–7 reference images alongside a text prompt.

curl -X POST "https://aiapi.fhddos.com/vidu/ent/v2/reference2video" \
  -H "Authorization: Bearer $TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "viduq1",
    "images": [
      "https://your.cdn.com/ref-1.jpg",
      "https://your.cdn.com/ref-2.jpg"
    ],
    "prompt": "A mouse runs toward the camera, smiling and blinking.",
    "duration": 5,
    "resolution": "1080p"
  }'

Reference-to-video exclusive parameters:

Parameter	Required	Description
`images`	✅	Array of 1–7 reference image URLs
`prompt`	✅	Text description, up to 2 000 characters

Creation response

All generation endpoints return the same initial response:

{
  "task_id": "176843862716480",
  "platform_id": "video_01JSGXXXXXXXXXXXXXXXXXX",
  "state": "created"
}

task_id is the raw Vidu upstream task ID; platform_id (video_<ULID>) is Fhddos’s internal tracking ID. Both IDs are accepted by the query and cancel endpoints. A created state means the task has been queued — poll the query endpoint to get the final result.

Task query & management

All video tasks are asynchronous. Poll the query endpoint or configure a callback_url on creation to receive state-change notifications.

Query task status

curl -X GET "https://aiapi.fhddos.com/vidu/ent/v2/tasks/$TASK_ID/creations" \
  -H "Authorization: Bearer $TOKEN"

Response example (completed):

{
  "id": "176843862716480",
  "state": "success",
  "credits": 10,
  "creations": [
    {
      "id": "creation_123",
      "url": "https://your.cdn/generate/abc.mp4",
      "cover_url": "https://your.cdn/generate/abc.jpg",
      "watermarked_url": "https://your.cdn/generate/abc_wm.mp4"
    }
  ]
}

Task states

State	Description
`created`	Task accepted and queued
`queueing`	Waiting for a generation slot
`processing`	Actively generating
`success`	Generation complete
`failed`	Generation failed (not charged)

Response fields

Field	Description
`creations[].url`	Temporary download URL for the generated video
`creations[].cover_url`	Cover thumbnail URL
`creations[].watermarked_url`	Watermarked video URL (only present when watermark was enabled)
`credits`	Credits consumed (deducted on success only)

Cancel a task

curl -X POST "https://aiapi.fhddos.com/vidu/ent/v2/tasks/$TASK_ID/cancel" \
  -H "Authorization: Bearer $TOKEN"

A successful cancel response includes any credits consumed up to that point.

Polling recommendations

Interval

Poll every 3–5 seconds to avoid rate limiting.

Timeout

Set a 5–10 minute client timeout; longer videos may take more time.

Callbacks

Set callback_url at creation to receive a push notification instead of polling.

Tips & best practices

Save 50 % with off-peak mode. Add "off_peak": true to any generation request. Generation may take slightly longer, but the cost is halved.

Add a watermark. Pass "watermark": true along with wm_position (position string) and wm_url (URL of your logo image) to brand every output video automatically.

Choose the right style. The style parameter on text-to-video accepts "general" for realistic/cinematic output and "anime" for stylised animation. Only Q3 models support the audio flag for simultaneous audio generation.

Model duration reference

Series	Supported durations
Q3 (`viduq3-*`)	1–16 seconds
Q2 (`viduq2-*`)	1–8 seconds
`vidu2.0`	4 s or 8 s
`vidu1.5`	Standard range; first-last frame not supported

​Authentication

​Core capabilities

​Available models

​Billing

​Quick start

​Text-to-video

​Query task status

​Video generation endpoints

​Common parameters

​Text-to-video

​Image-to-video

​First-last frame

​Reference-to-video

​Creation response

​Task query & management

​Query task status

​Task states

​Response fields

​Cancel a task

​Polling recommendations

Interval

Timeout

Callbacks

​Tips & best practices

​Model duration reference

Authentication

Core capabilities

Available models

Billing

Quick start

Text-to-video

Query task status

Video generation endpoints

Common parameters

Text-to-video

Image-to-video

First-last frame

Reference-to-video

Creation response

Task query & management

Query task status

Task states

Response fields

Cancel a task

Polling recommendations

Tips & best practices

Model duration reference