Multi-view reference rendering via the Wan 2.7 API lets us spin up character animation reels in minutes — no rigging, no manual matching across angles.
The Wan 2.7 API renders a 720p or 1080p video from multi-view reference images plus a natural-language prompt (Alibaba Wan 2.7). The Wan 2.7 API preserves subject identity across the reference views, supports 2-to-15-second outputs at 30 FPS, seed lock, enable_prompt_expansion auto-enrichment, and negative prompts — all through one REST endpoint, metered per second at the official rate.
Input
The text prompt sent to the model
Output
SampleRun the model to replace this previewWan 2.7 API: Reference to Video
Use the Wan 2.7 API to render a 720p or 1080p clip from multi-view reference images and a natural-language prompt.
The Wan 2.7 API renders a 720p or 1080p video clip from multi-view reference images plus a natural-language prompt (Alibaba Wan 2.7). The Wan 2.7 API preserves subject identity across the reference views — 2-to-15-second outputs at 30 FPS, seed lock, prompt-expansion auto-enrichment, and negative prompts — all through one REST endpoint.
The Wan 2.7 API is metered per second at the official rate: $0.10/s at 720p, $0.15/s at 1080p — $0.50 per 5-second 720p clip (default), up to $1.50 at 10s/1080p. One RouterBase key works alongside 200+ other models — no provider-specific SDK.
Six reasons teams ship on the Wan 2.7 API
From multi-view identity locking to prompt-driven motion — what the Wan 2.7 API gets right.
Multi-view references
The Wan 2.7 API takes multiple reference images of the same subject from different angles and locks identity across the rendered clip — no drift between views.
Natural-language motion
Describe the action in plain English; the Wan 2.7 API renders a 2-to-15-second clip from the multi-view references — no storyboard, no timeline editing.
Identity consistency
Subject face, styling, and silhouette stay anchored to the reference views. The Wan 2.7 API blends views internally for stable identity rendering.
720p / 1080p, 30 FPS
Switch resolution per Wan 2.7 API call. Default 720p for fast iteration; 1080p for delivery. Aspect ratio is set by the reference images.
Seed + prompt expansion
Lock outputs with seed for exact reproduction. Toggle enable_prompt_expansion and the Wan 2.7 API auto-enriches your prompt before generation.
Metered per second
The Wan 2.7 API charges at the official per-second rate — $0.10 at 720p, $0.15 at 1080p. Pay only for the seconds you render.
Quickstart
Get started with the Wan 2.7 API in 3 steps
From sign-up to your first Wan 2.7 API call in under 5 minutes.
Create a RouterBase API key
Sign up and generate sk-rb-… — one key works across the Wan 2.7 API and 200+ other models on RouterBase.
POST the reference images
Send the multi-view reference images plus prompt in JSON to the Wan 2.7 API endpoint. The call is async — RouterBase polls upstream until the render completes (typically 30 s–3 minutes), then returns a signed URL.
Inspect the result
Each Wan 2.7 API response includes generation metadata for auditing and a render URL ready to embed in your product.
Pay only for what you use
RouterBase passes through partner-tier pricing. Compared against the model's official published API rate.
| Resolution | RouterBase |
|---|---|
| 1280*720 / 720*1280 / 960*960 / 1088*832 / 832*1088 5s | $0.500 |
| 1920*1080 / 1080*1920 / 1440*1440 / 1632*1248 / 1248*1632 5s | $0.750 |
About $0.150 per second.One run at the default settings (1920*1080 / 1080*1920 / 1440*1440 / 1632*1248 / 1248*1632) costs $0.750.With $1 you can run this model approximately 1 times.
What teams build with the Wan 2.7 API
Real production loads running on the RouterBase model catalog across 200+ models.
We swapped three providers through RouterBase in one line of code when OpenAI had its outage. Our users never noticed.
The Wan 2.7 API holds identity across our 500-portrait talent catalog — the same person renders from any view, render after render.
One Wan 2.7 API call ingests three reference angles and outputs a unified motion clip — we collapsed a multi-stage pipeline into a single endpoint.
Plugged the Wan 2.7 API into my existing OpenAI SDK with one base URL swap — less work than writing this testimonial.
RouterBase puts 200+ models behind one key. Our research team stopped asking for new provider accounts.
Seed lock on the Wan 2.7 API gives us deterministic outputs for regression testing — same prompt, same seed, same references, same clip every time.
We migrated 42 services to RouterBase over a weekend. The only PR comment we got was 'wait, that's all?'.
Per-second metering on the Wan 2.7 API means we pay only for the seconds we render — at our scale that's cleaner budgeting than per-clip pricing.
Frequently Asked Questions
Common questions about the Wan 2.7 API.
The Wan 2.7 API is RouterBase's pass-through to Alibaba Wan 2.7 reference-to-video. It renders a 720p or 1080p clip from multi-view reference images plus a natural-language prompt — 2–15 second durations at 30 FPS, identity consistency across views, seed lock, prompt-expansion, and negative prompts.