pixverse-v6

Quick example

curl https://reapi.ai/api/v1/videos/generations \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "pixverse-v6",
    "prompt": "A neon-lit Tokyo alley at night, light rain, anamorphic flare",
    "size": "21:9",
    "resolution": "720p",
    "duration": 8
  }'

import requests

resp = requests.post(
    "https://reapi.ai/api/v1/videos/generations",
    headers={
        "Authorization": "Bearer YOUR_API_KEY",
        "Content-Type": "application/json",
    },
    json={
        "model": "pixverse-v6",
        "prompt": "A neon-lit Tokyo alley at night, light rain, anamorphic flare",
        "size": "21:9",
        "resolution": "720p",
        "duration": 8,
    },
    timeout=30,
)
print(resp.json())

const r = await fetch("https://reapi.ai/api/v1/videos/generations", {
  method: "POST",
  headers: {
    Authorization: "Bearer YOUR_API_KEY",
    "Content-Type": "application/json",
  },
  body: JSON.stringify({
    model: "pixverse-v6",
    prompt: "A neon-lit Tokyo alley at night, light rain, anamorphic flare",
    size: "21:9",
    resolution: "720p",
    duration: 8,
  }),
});
console.log(await r.json());

package main

import (
    "bytes"
    "encoding/json"
    "fmt"
    "io"
    "net/http"
)

func main() {
    body, _ := json.Marshal(map[string]any{
        "model":      "pixverse-v6",
        "prompt":     "A neon-lit Tokyo alley at night, light rain, anamorphic flare",
        "size":       "21:9",
        "resolution": "720p",
        "duration":   8,
    })
    req, _ := http.NewRequest("POST",
        "https://reapi.ai/api/v1/videos/generations", bytes.NewReader(body))
    req.Header.Set("Authorization", "Bearer YOUR_API_KEY")
    req.Header.Set("Content-Type", "application/json")

    resp, _ := http.DefaultClient.Do(req)
    defer resp.Body.Close()
    out, _ := io.ReadAll(resp.Body)
    fmt.Println(string(out))
}

Submit response

{
  "id": "task_018f5a3a1b6e7d9f8c2b4d6e8f0a2c4e",
  "model": "pixverse-v6",
  "status": "processing",
  "created_at": 1735000000
}

Poll GET /api/v1/tasks/{id} (see the Tasks reference) until status === "completed". The completed payload's output.video_urls holds the generated MP4 URL, valid for 7 days.

Authentication

Every call needs a Bearer token. Generate keys at reapi.ai/settings/apikeys.

Authorization: Bearer YOUR_API_KEY

Keys carry the active workspace's billing scope — there is no separate project header.

Endpoint

POST /api/v1/videos/generations
GET  /api/v1/tasks/{id}

Submission is async. The POST returns immediately with a task_id; the task endpoint returns the same envelope until completion. Polling does not consume credits.

Mode routing

pixverse-v6 picks its mode from which fields you send — there is no mode parameter. Matching is by priority; the first satisfied condition wins:

Request body

model — required

string. Always "pixverse-v6".

prompt — string, required

Up to 5,000 characters. Required in every mode. Empty / whitespace-only prompts are treated as missing.

resolution — string, default "540p"

One of 360p / 540p / 720p / 1080p. Drives pricing.

duration — integer, default 5

Output length in seconds. Any integer in [1, 15]. First/Last-frame mode accepts only 5 or 8. Drives pricing linearly: ceil(per_second_usd × duration × 1000) credits (1 credit = $0.001).

size — string, default "16:9"

Output aspect ratio. Honored in text-to-video and multi-reference fusion only (other modes derive the ratio from the source media). One of:

seed — integer, default 0

Reproducibility hint. Range [0, 2147483647]. Same prompt plus seed returns a similar (not bit-for-bit identical) result.

negative_prompt — string

Up to 2,048 characters. Describes content to keep out of the frame.

audio — boolean, default false

When true, the model synthesizes a soundtrack alongside the video. Audio raises the per-second rate (see Pricing). Default false is silent.

watermark — boolean, default false

When true, adds a watermark to the bottom-right corner.

motion_mode — string

pixverse-v6 supports only "normal". Other values are rejected upstream.

generate_multi_clip_switch — boolean, default false

When true, generates a multi-segment continuous video. Supported in text-to-video and image-to-video only.

image_urls — string[] (image-to-video)

Public HTTP(S) URL array. The model uses the first entry as the starting frame.

first_frame_image / last_frame_image — string (first/last-frame)

Public HTTP(S) URLs. Must be sent as a pair; the model interpolates motion from the first frame to the last over a 5 or 8 second clip.

img_references — string[] (multi-reference fusion)

1 to 7 public HTTP(S) URLs. The model fuses subjects from each reference into one clip.

extend_from_task_id — string (video extend)

The reAPI task id (task_...) of one of your own completed pixverse-v6 videos. Do not pass the upstream provider task id — reAPI resolves that internally before sending the extend request. The model continues the scene from where that clip ended.

Response envelope

Submit and poll share the same shape — only status and output fill in over time.

{
  "id": "task_018f5a3a1b6e7d9f8c2b4d6e8f0a2c4e",
  "model": "pixverse-v6",
  "status": "completed",
  "created_at": 1735000000,
  "output": {
    "video_urls": ["https://cdn.reapi.ai/media/tasks/018f5a3a1b6e7d9f8c2b4d6e8f0a2c4e/0.mp4"]
  },
  "error": null
}

output.video_urls URLs are valid for 7 days. Re-host to your own storage if you need them longer. To extend a clip later, keep the task id — pass it as extend_from_task_id.

Validation errors

All cases below return HTTP 400. Pattern-match on code, not message — message strings carry request-specific context (field names, observed values) and are not a stable contract.

The full envelope is { "error": { "code", "message", "request_id" } } — see Errors catalog for the wire format and request_id correlation tips.

Recipes

Text-to-video — minimum request

{
  "model": "pixverse-v6",
  "prompt": "A corgi running through a sunflower field at golden hour"
}

Text-to-video — full parameters

{
  "model": "pixverse-v6",
  "prompt": "A neon-lit Tokyo alley at night, light rain, anamorphic lens flare",
  "size": "21:9",
  "resolution": "720p",
  "duration": 8,
  "seed": 42,
  "audio": true
}

Image-to-video — animate a single frame

{
  "model": "pixverse-v6",
  "prompt": "Camera slowly zooms in, gentle wind moves the leaves",
  "image_urls": ["https://your-cdn.com/first_frame.jpg"],
  "resolution": "540p",
  "duration": 5
}

First/Last-frame transition

{
  "model": "pixverse-v6",
  "prompt": "Transform smoothly from a puppy to a cat",
  "first_frame_image": "https://your-cdn.com/puppy.jpg",
  "last_frame_image": "https://your-cdn.com/cat.jpg",
  "resolution": "540p",
  "duration": 5
}

Multi-reference fusion

{
  "model": "pixverse-v6",
  "prompt": "A girl wearing the outfit from image 2, holding the cat from image 3",
  "img_references": [
    "https://your-cdn.com/character.jpg",
    "https://your-cdn.com/outfit.jpg",
    "https://your-cdn.com/cat.jpg"
  ],
  "size": "9:16",
  "resolution": "720p",
  "duration": 5
}

Video extend

{
  "model": "pixverse-v6",
  "prompt": "The character now walks into a forest",
  "extend_from_task_id": "task_018f5a3a1b6e7d9f8c2b4d6e8f0a2c4e",
  "resolution": "540p",
  "duration": 5
}

Choosing a mode

Polling pattern

The task endpoint behaves identically to other video tasks — the completed output carries video_urls. A pragmatic schedule:

0–5 minutes:    poll every 5s
5 min – 1 h:    back off gradually toward 1 min
≥ 1 h:          cap at 3 min between polls

A typical task completes in a few minutes.

Pricing

Per-second × resolution. Turning audio on raises the per-second rate at every tier; every other parameter is free. See current rates on the Pixverse v6 model page.

Bill formula (1 credit = $0.001):

credits = ceil(per_second_usd × duration × 1000)

Failed jobs refund automatically.

Tips

• Prompt motion, not just scene. "Slow push-in, warm tones, shallow depth of field" outperforms a pure noun-list of what's on screen.
• Keep references on-topic. Multi-reference fusion works best when each image contributes a distinct, clearly-framed subject (a face, an outfit, a prop) rather than overlapping whole scenes.
• Two frames need to make sense as endpoints. First/Last interpolation is smoothest when the two frames share enough subject and composition.
• Chain extends one beat at a time. Build a longer sequence by extending the latest clip's task id repeatedly instead of asking for one long shot.
• Audio is opt-in here. Leave audio off unless you need a soundtrack — it raises the per-second rate.

Related

• Errors catalog
• Authentication
• Quickstart