Try MiniMax FastH3 V2 from Fastvideo →
Models
Agents
Workflows
Studio
PricingBlogDocs
ExploreDiscover models by categoryBrowse All ModelsBrowse the complete catalogSee FavoritesSign in to view saved models
Generative Media AgentCreate and edit media by chattingWorkflow AgentBuild visual workflows with Agent
OverviewThe platform at a glanceLearnSkills, knowledge, guardrailsAnatomyWhat makes agents reasonBuild Your AgentPick skills, set tier, deploy
Pre-built AgentsBrowse the catalog
Agent Usecases
Ad Campaign ManagerApp Event ManagerApp Review RepliesBarber BookingCustomer Win-BackEcommerce ListingsRestaurant Reviews
Sign InStart Building

Task History

Click to see output list

No tasks yet

Go to Models
Explore models/
Video GenerationActive

Pruna P-Video 2 Pro Reference

bypruna

Generate 5–15s videos from a prompt plus reference images, clips, and audio. Pruna’s P-Video 2 Pro Reference outputs 480p/768p MP4 with sound.

Image to Video
Model ID
p-video-2-pro-reference
Provider
pruna
Updated
1792080060
5
Comments
Average rating : 3.33 (9 users)
Providerpruna
Modelp-video-2-pro-reference
Image to Video
Support Wiro on G2
wiro playground—pruna/p-video-2-pro-reference
Reset to defaults
0 / 1
Maximum 1 image allowed

Optional: Soft start keyframe. Counts as a reference image

Required: Describe the video to generate and how the reference images, videos and audios should be used

Required: Length of the generated video

Required: Output resolution. The video is always 24 fps and includes generated audio

Required: Generation recipe. Speed gives better results, cost is cheaper

Optional: Output aspect ratio. Adaptive follows the references

Sample outputs
Updated 1792080060

Overview

P-Video 2 Pro Reference is a reference-guided video generator by Pruna. It follows your written scene prompt, then uses your images, short videos, and optional audio as consistency anchors. It produces a single short clip with sound, so you can keep a character, product, or setting stable across shots without rebuilding the look each time.

This model is built on MiniMax H3. It’s designed for “reference-to-video” work. Your references guide identity, style, motion, and audio feel. They do not automatically become the first frame.

What you can build

  • Character-consistent clips for ads, reels, and short cinematic beats
  • Product videos that keep the same packaging shape and materials across takes
  • Location continuity shots that match lighting, palette, and set dressing from references
  • Motion studies where a reference clip guides camera framing and movement
  • Audio-led scenes where a reference track steers dialogue timing, music feel, or pacing

Inputs

  • A written prompt that describes what the clip should show, including action and camera direction, plus any sound you want in the result.
  • At least 1 reference image or reference video.
  • Optional reference images (up to 9 total images). This total includes any optional opening and closing keyframe images. Each image must be 20 MiB or smaller and 4096 pixels or smaller on each side.
  • Optional reference videos (up to 3 files). Each video must be 2–15 seconds. All reference videos combined can total 15 seconds. Each video must be 100 MiB or smaller, up to 1080p, and up to 60 fps. Longer videos may be trimmed.
  • Optional reference audio (up to 3 files). All reference audio combined can total 15 seconds. Each audio file must be 50 MiB or smaller. Longer audio may be trimmed.
  • Optional opening keyframe image to softly steer the first frame. This counts toward the 9-image limit.
  • Optional closing keyframe image to softly steer the last frame. This counts toward the 9-image limit.
  • A target duration from 5 to 15 seconds.
  • An output resolution choice of 480p or 768p.
  • An aspect ratio choice. You can also choose an adaptive setting that follows the references.
  • An optional prompt rewrite strength. It can expand your prompt before generation.
  • An optional seed value when you need repeatable runs from the same inputs.

Outputs

  • One MP4 video clip.
  • The clip is 5–15 seconds long.
  • The clip runs at 24 fps.
  • The clip includes an audio track generated with the video.

Limitations

  • You can’t run it with a prompt alone. You must provide at least one reference image or reference video.
  • The opening and closing keyframe images are soft controls. They guide the ends, but they don’t guarantee exact frame matches.
  • The model enforces strict reference caps. Up to 9 images total, 3 videos total, and 3 audio files total.
  • Reference videos must stay within the length limits. Longer clips may be trimmed.
  • There’s no quality recipe for this model. It supports only the speed and cost recipes.
  • Low-quality references can reduce consistency. Avoid blurry images, heavy compression, extreme crops, and mismatched lighting across references.
  • This model generates new footage. It is not a video editor for modifying an existing clip.

Safety & compliance

  • Follow Pruna’s terms and acceptable use rules when generating or sharing outputs.
  • This model is subject to the MiniMax H3 Community License and its acceptable use policy.
  • Don’t use the model or its outputs to train or improve other AI models.
  • Only upload reference media you have rights to use. This includes faces, brands, and copyrighted content.

Example prompts

Great starting points for p-video-2-pro-reference.

Turquoise waves from the reference image roll in one after another, curl and break into white foam that rushes up the wet sand and slowly pulls back, leaving shimmering reflections of the golden sky. Palm leaves sway gently in the breeze, the camera slowly glides forward along the shoreline, crashing waves and soft windVideo Generation
The flamenco dancer from the reference image raises her arms, spins sharply and sweeps the ruffles of her dress, then strikes the wooden floor with rapid heel stamps. The spotlight follows her, fast Spanish guitar strumming, handclaps and the sound of her heelsVideo Generation
The street artist from the reference image shakes a spray can, then sweeps it across the wall, adding bright orange and electric blue strokes to the tiger mural as fine paint mist drifts through the streetlight. The camera slowly arcs around him to reveal the mural, the rattle of the can, hissing spray and distant city trafficVideo Generation
The origami crane from the reference image slowly comes to life, flaps its paper wings with crisp folding sounds and lifts off the book, then circles the sunlit desk as the book's pages flutter in the draft. It flies toward the bright window and glides along the glass, gentle piano melodyVideo Generation

API quick start

Run p-video-2-pro-reference with a single API call.

POST https://api.wiro.ai/v1/Run/pruna/p-video-2-pro-reference
{
  "prompt": "Turquoise waves from the reference image …",
  "duration": 5,
  "resolution": "768p",
  "mode": "speed"
}
curl
curl -X POST "https://api.wiro.ai/v1/Run/pruna/p-video-2-pro-reference" \
  -H "Content-Type: application/json" \
  -H "x-api-key: YOUR_WIRO_API_KEY" \
  --data-binary @- <<'JSON'
{
  "prompt": "Turquoise waves from the reference image …",
  "duration": 5,
  "resolution": "768p",
  "mode": "speed"
}
JSON
View full API docs

Discover, test, and run AI models, build workflows and agents with one unified API.

All systems operational
WiroAboutBlogCareersContact
ProductModelsAgentsPricingPartnerChangelogStatusFAQ
ModelsNano Banana 2GPT Image 2.5Seedream V5 ProSeedance 2.5Veo 3.1Kling V3FLUX 3FLUX.2 ProWan 3.0 PrimeGrok Imagine 1.5
PartnersGoogleOpenAIByteDanceBlack Forest LabsKling AIQwenAlibabaxAIMiniMaxElevenLabs
Getting StartedIntroductionAuthenticationProjectsCode ExamplesWiro MCP ServerSelf-Hosted MCPn8n IntegrationLLMs.txt
API ReferenceModelsRun a ModelModel ParametersTasksLLM & Chat StreamingWebSocketRealtime VoiceFiles
© 2026 Wiro AI. All rights reserved.
PrivacyTermsData Deletion