P-Video-Edit
Pruna AI's hosted video editor, drafts included
- referenceImages
- video
P-Video-Edit is Pruna AI's video editing model - the video sibling of P-Image, the speed-first company behind fast-FLUX work like FLUX-Juiced. Where P-Image made its name on "four images in six seconds," the video line is aimed at the editing side of the job: take a clip you already have, describe what changes, get a new version back. And this node's design shows the discipline that makes hosted video editing financially survivable: it has a draft mode, and you should use it.
The core mechanic is prompt-driven video editing, which is a different creature from text-to-video. Both required inputs make that clear: video takes your source clip as a UUID or URL, and positivePrompt describes the edit. The output follows the source's aspect ratio and duration, always at 25 fps, with the output size fixed per aspect ratio - the tooltip is upfront that the model won't re-time or re-frame your clip, it re-renders it. That constraint is actually a feature for editing workflows: no surprises about length or orientation, which is exactly what you want when the clip has to slot back into an edit. referenceImages can supply a look or subject reference alongside.
The rest of the controls are where Pruna's cost-consciousness shows. settings.draft (default off) gives you a faster, lower-quality preview billed at the draft rate - the output resolution is unchanged, so you're really paying for speed while validating the edit. That's the single most important control on the node, because it's your defense against paying full video rates for an edit that misses. settings.promptUpsampling defaults on to auto-enhance your prompt; settings.audio defaults on to keep audio in the output. seed, numberResults, the safety group with fast/full modes, outputFormat (MP4/WEBM/MOV), and outputQuality complete the list. Output: VIDEO.
Install
The usual Runware pack setup covers it: ComfyUI Manager → search Runware → install → restart, or clone https://github.com/Runware/ComfyUI-Runware into custom_nodes and pip install -r requirements.txt. Runware API key from the dashboard - Settings → Runware API key, RUNWARE_API_KEY, or runware auth login.
The honest workflow
This node lives or dies on its draft loop. Validate the edit in draft mode, then spend full-rate only on the version you're keeping - and read the per-run cost on the node's title bar like a meter. Two caveats from the shape of the thing: P-Video-Edit is a closed hosted model (Pruna's video line isn't open-weights), so there's no local fallback and no community fine-tuning scene to lean on; and your source clips leave your machine through Runware's cloud and moderation. If the footage is proprietary, that's the dealbreaker to weigh before you queue.
Inputs (15)
| Name | Type | Default | Description |
|---|---|---|---|
| video | STRING | Video input (UUID or URL). The output follows the source aspect ratio and duration, always at 25 fps. Output size is fixed per aspect ratio and never exceeds 848 pixels on the long edge, regardless of the source resolution. | |
| positivePrompt | STRING | Text prompt describing elements to include in the generated output. | |
| referenceImagesopt | IMAGE | — | |
| seedopt | INT | 00–2147483647 | Random seed for reproducible generation. When not provided, a random seed is generated in the unsigned 32-bit range. |
| numberResultsopt | INT | 11–4 | Number of results to generate. Each result uses a different seed, producing variations of the same parameters. |
| settings.audioopt | BOOLEAN | true | Include audio in the generated video output. |
| settings.draftopt | BOOLEAN | false | Enable draft mode for a faster, lower-quality preview billed at the draft rate. The output resolution is unchanged. |
| settings.promptUpsamplingopt | BOOLEAN | true | Automatically enhance the prompt for improved results. |
| safetyopt | BOOLEAN | false | Enable to set safety. Off uses the model's default. |
| safety.checkContentopt | BOOLEAN | false | Enable or disable content safety checking. Increases total generation time. |
| safety.modeopt | COMBO | fast | Safety checking mode for video generation. |
| ttlopt | BOOLEAN | false | Enable to set ttl. Off uses the model's default. |
| ttl_valueopt | INT | 60 | Time-to-live (TTL) in seconds for generated content. Only applies when `outputType` is `URL`. |
| outputFormatopt | COMBO | MP4 | File format for the generated video. |
| outputQualityopt | INT | 9520–99 | Compression quality of the output. Higher values preserve quality but increase file size. |
Outputs (1)
| Name | Type | Description |
|---|---|---|
| video | VIDEO | — |