A REST API for generating AI videos from text prompts or images — with effects, transitions, lip-sync, and sound — from the PixVerse platform.
All endpoints are under https://app-api.pixverse.ai/openapi/v2/. Every request uses two headers: API-KEY (your key) and Ai-trace-id (a fresh UUID per request — reusing the same value returns the previous result instead of triggering a new generation). All responses use the shape { ErrCode, ErrMsg, Resp } — ErrCode=0 means success and your payload is in Resp. Video generation is asynchronous: submit a job, receive a video_id, then poll the get-status endpoint every 3–5s until status=1 (success). Shared parameters (model, duration, quality) — including valid enum values — are documented inline on each endpoint page. API credits are billed separately from the PixVerse web app and purchased on the Billing page.
Introduction to PixVerse Platform: What the platform generates — text-to-video, image-to-video, transitions, effects, lip-sync, sound — and how the API exposes it.
Model & Pricing: Credit cost per operation by model and feature, so you can estimate cost before calling anything.
Rate limit: Per-account request and concurrency limits — check this before sizing your integration.
Quick Start: Shortest path from zero to first successful video — quick overview before the full walkthrough below.
How does the API work?: The full async pattern in one page — API-KEY + Ai-trace-id headers, submit a generation job, receive a video_id, poll the status endpoint until status=1.
How to get API key?: Step 1 — create an account and generate your API key from the dashboard (shown once, save it immediately).
How to subscribe API plans?: Step 2 — buy credits on the Billing page; note API credits are separate from the PixVerse web app membership.
Upload Image: Step 3 — POST multipart form-data with an image field (JPG/PNG/WebP, max 20MB and max 10000px; URL-based uploads are not supported). Returns Resp.img_id (integer) to reference in the next call.
Image-to-Video Generation: Step 4 — POST JSON body with fields img_id, prompt, model, duration, quality, aspect_ratio, motion_mode, negative_prompt. Verified sample values: model="v6", duration=5, quality="540p", aspect_ratio="16:9", motion_mode="normal". Returns Resp.video_id.
Get Video Generation Status: Step 5 — GET with the video_id every 3–5s until Resp.status=1. Other statuses: 5 = generating, 6 = deleted, 7 = moderation failed, 8 = generation failed. On success Resp.url contains the final video URL.
Multi-transition Video Generation: Generate a longer video that transitions through multiple keyframe images in sequence (2–7 keyframes, 1–30s output).
Extend Generation: Extend an existing generated video with a continuation, producing a longer clip from a previous video_id.
Fusion (Reference to Video) Generation: POST /openapi/v2/video/fusion/generate with an image_references array (each: type: "subject"|"background", img_id, ref_name) plus a prompt that uses @ref_name to place subjects/backgrounds in the scene. Requires model v4.5+.
Swap Video Generation: POST /openapi/v2/video/swap/generate — replace a selected object in an existing video with a target image. Needs video_media_id, keyframe_id, mask_id, img_id; pair with the mask/selection API to specify the replacement region.
Restyle Video Generation: Re-render an existing video in a different visual style (3D, anime, cinematic, enhanced realism, etc.).
Restyle effect list: Catalog of valid style presets to pass to the Restyle endpoint.
Video Effects Generation: PixVerse's signature template-based effects (character transforms, cinematic motion, etc.) — driven by passing template_id + prompt to text-to-video or image-to-video. Template catalog lives in the web Effect Center; multi-image effects take img_ids instead of img_id.
Mimic Generation: Transfer motion from a reference video onto a character image — inputs are a character image + a reference video (mp4/mov, max 1920p, 100MB, 30s). For choreography and animation reuse.
Modify Generation:edit any part of an existing video — including adding, replacing, removing, modifying elements, or transforming the overall style.
Image Template Generation: Image-centric template workflow — upload photos, pick a template from the Template Center, get AI-composed image outputs.
