P-Video 2 Pro
Pruna AI's quick video model with a soundtrack and legible lettering, the default for paid accounts in NeverNo's Video Studio, reached without your name or IP.
Updated November 9, 2026
A humpback whale gliding through golden clouds above a 1950s seaside town at dusk
At a glance
| Maker | Pruna AI |
|---|---|
| Strengths in the app | Fast, Sound |
| Arena Elo (Artificial Analysis) | 1,128 |
| Typical wait | about 16 s |
| Who can use it | Paid plans |
| How it’s paid | Credits, from the first use |
| Privacy | Anonymized: NeverNo sends the request to its maker without your name, account or IP address. The maker still reads the text, applies its own rules and may keep it for a while under its own policy. |
Price
What the app charges, read from the same table the servers bill with. 1 credit = $0.01.
| 5-second clip | 47 credits ($0.47) |
|---|---|
| 10-second clip | 93 credits ($0.93) |
What it is
P-Video 2 Pro is a video model from Pruna AI, and it is the model a paid NeverNo account starts with in the Video Studio. It makes clips of 5 or 10 seconds in 16:9 or vertical 9:16, at 1344 by 768 pixels, from a text description or from a starting image with an optional ending image. The Studio tags it fast and audio: the clip comes with its own sound in the same pass. Video is for paid plans and each clip is paid in credits.
The headline is speed. A typical 5 second clip takes about 16 seconds to render, as measured on Runware, our video provider, against 33 seconds for HeyGen Video 1, 36 for Veo 3.1 Fast, 49 for LTX 2.3 Fast and well over two minutes for Wan 3.0 and Seedance 2.5. In our test it also wrote a shop sign legibly and kept a boat whole, two things LTX 2.3 Fast got wrong, which is why the Studio tags it for legible text.
On the Artificial Analysis video arena, read on November 7, 2026, P-Video 2 Pro has an Elo of 1128 from text and 1117 from an image. MiniMax H3 Max reaches 1145 and 1115, HeyGen Video 1 1152 and 1123, Wan 3.0 1172 and 1109, and Seedance 2.5 1155 and 1158. So the slower models edge ahead on paper, while this one answers in a fraction of the time, which is exactly why it is the sensible default.
It is anonymized. NeverNo sends the request to Pruna AI without your name, account or IP address; the lab still reads the prompt and applies its own rules, and some requests it declines. The finished clip is stored in your browser, not on our servers. Its rhythm suits a loop of experiments. Try one scene at 5 seconds first, since a short clip reveals problems with motion, framing or sound immediately, then repeat the winning wording at 10 seconds once you are sure. Keep a note of the phrases that worked, such as a camera move, a time of day or an ambient noise, and reuse them as building blocks. Paid accounts start here for a reason: the quick answer lets you iterate honestly, and the sound track spares you a second tool for every idea you try. Small agencies, teachers and shop owners often need a presentable clip by the end of a meeting, and that is the situation where a sixteen second wait, rather than a coffee break, makes the difference. Sort your saved clips into folders by project, and a quick pass in your editor will show which takes deserve a longer, slower render elsewhere. Scrub through the finished clip frame by frame once, looking for flicker, warped hands or drifting letters, before you decide a take is final.
Strengths
About 16 seconds for a typical 5 second clip, the shortest wait among the Studio's paid video models.
Sound comes with the picture, so you do not need a separate step for ambience.
Tagged for legible text, useful for shop signs, titles and labels in the frame.
Start image plus optional end image, in 16:9 or 9:16, at 1344 by 768 pixels.
Arena Elo of 1128 from text and 1117 from an image (Artificial Analysis, read November 7, 2026).
Limits
Slightly below the slowest models on the arena: Wan 3.0 and Seedance 2.5 score higher for cinematic scenes.
Pruna AI's content policy is not published, so some prompts may be declined; the model is anonymized, not private.
Needs a paid plan, and every clip is paid in credits.
When to choose it
Quick social clips with sound
Describe a scene, pick 9:16, and get a clip with its own ambience in well under a minute. Good for shop windows, recipe teasers and event announcements where the sound matters as much as the picture.
Animating a product photo
Give a still of a bottle, a sneaker or a poster as the start image, add a camera move and a sentence about the light, and obtain a short product loop. Add an end image to control where the shot lands.
Trying several versions of one idea
At this speed you can render three or four takes of the same prompt in a few minutes, keep the best, and only then spend credits on a slower, higher-rated model for the final.
Prompt tips
Say what the sound should be as well as what the picture shows, for example waves on a pier or a crowd murmuring in a market.
Write signs and titles inside quotation marks, short and in plain words, and keep them to one object in the scene.
Choose a camera move from the Studio's list instead of describing it at length; one sentence on the movement and one on the light is enough.
Frequently asked questions
Does P-Video 2 Pro make sound?
Yes. The Studio tags it audio, and the clip is made with its soundtrack in the same pass.
How fast is P-Video 2 Pro?
A typical 5 second clip takes about 16 seconds, measured on Runware, our video provider.
Is it private?
It is anonymized: NeverNo sends the request without your name, account or IP address, but Pruna AI still reads the prompt and applies its own rules. Clips are stored in your browser.
Can I start from an image?
Yes. It accepts a starting image and an optional ending image, in 16:9 or 9:16.
P-Video 2 Pro or MiniMax H3 Max?
MiniMax H3 Max scores higher from text (1145 versus 1128, read November 7, 2026) and adds a cinematic tag, but P-Video 2 Pro is the default of paid accounts and is faster to try ideas on.