Kuponaut
CodesComparisonsGuidesCalculatorAboutContactSubmit a couponTR

Kuponaut › Blog

How to turn a photo or a prompt into an AI video with PixVerse (with sound)

PixVerse turns a written prompt or a still photo into a short video clip, and its V6 model can generate the sound at the same time. This guide covers choosing between text-to-video and image-to-video, writing a prompt the model can follow, the V6 settings PixVerse documents, what credits and the free plan mean in practice, and how to label the result when you post it. Feature details come from PixVerse's model pages, product pages and blog; prices are from its subscription page on 15 October 2026.

By Fırat Papatya · Published · 7 min read

Affiliate disclosure: links to PixVerse in this article are partner links. Kuponaut does not have a PixVerse code yet; we have asked for one. If you buy through our links, Kuponaut may earn a commission at no extra cost to you. Prices were read on app.pixverse.ai on 15 October 2026 and can change.

Text-to-video or image-to-video

PixVerse offers both routes on the web and in its iOS and Android apps. Text-to-video starts from a written idea. Image-to-video starts from a picture you upload, such as a product shot, a portrait, an illustration or a concept frame. PixVerse's image-to-video page says the image route is often faster when you already have a reference, because the model does not have to invent how the subject looks.

Use text-to-video for new concepts where any good-looking result will do. Use image-to-video when the look matters: your product, your character design, a real place. PixVerse recommends images with a clear subject, readable edges and strong composition. Only upload photos you have the right to use, and do not animate real people who have not agreed to it.

Step 1: choose V6 and set the format

PixVerse V6, released on 30 March 2026, is the company's flagship model. PixVerse describes it as "1080p Video with Native Audio": it generates dialogue, sound effects and ambient sound together with the picture, instead of leaving sound for a separate editing step. The site lists other models too, including PixVerse R2, Seedance 2.5, Seedance 2 and MiniMax H3, but on its image-to-video page native audio is listed only for V6.

PixVerse's V6 guide says to set the output before writing the final prompt. Pick PixVerse V6 in the model or workflow selector, then choose text-to-video or image-to-video. Then set the duration, quality, aspect ratio and audio setting. PixVerse's V6 review lists durations from 1 to 15 seconds, quality of 360p, 540p, 720p or 1080p, and aspect ratios of 16:9, 1:1, 9:16, 3:2, 2:3 and 21:9 in supported workflows.

Choose the ratio for where the clip will go: 9:16 for TikTok, Reels and Shorts, 16:9 for YouTube or a website, 1:1 for feed posts. PixVerse suggests drafting at lower quality, which costs less, and keeping 1080p for a final version you have already checked. Your plan also caps the resolution: the plan table lists 540p for the free plan and 720p for Standard.

Step 2: write the prompt: subject, motion, camera, sound

PixVerse's text-to-video page gives a formula: subject + setting + action + camera + light + sound + format. It says prompts should read like short directing notes. The V6 review recommends literal, descriptive wording over metaphors, and warns against vague adjectives with nothing visible attached to them.

Subject: say who or what is in the shot and what they look like. Motion: describe one clear action rather than several. Camera: name the movement. The V6 page lists slow push-in, handheld tracking shot, low-angle shot, aerial view, POV movement, pan, tilt, orbit and close-up, and says the clearer the camera language, the easier it is for the model to follow. PixVerse's review describes no separate camera panel, so camera direction goes in the prompt.

Sound: describe it explicitly, because V6 builds the audio from your words. PixVerse's examples use lines such as "Sound: rolling surf, howling wind". For dialogue, write the line and the tone of voice. A prompt in this format might read: "A woman in a yellow raincoat walks along a harbour wall at dusk, waves breaking behind her. Slow tracking shot from the side, low angle. Warm orange light. Sound: waves, gulls, wind. She says quietly: 'We made it.' 9:16."

For image-to-video, PixVerse's sample prompt first states that the uploaded image is the reference and first frame, then describes small, specific movements, sets the camera (a subtle push-in) and lists exclusions such as "no new animals, no people, no object duplication, no morphing". Exclusions like these help stop the model adding things that were not in your photo.

Step 3: generate, review and retry

Click Create and watch the result closely. PixVerse lists what to check: character consistency, motion stability, lens behaviour and sound timing. It adds that dialogue, pronunciation and subtitle accuracy still need a human check, and its V6 launch announcement says precise directional control in complex scenes is still evolving.

When a take is close, change one thing per retry, such as the camera move, one phrase or the duration, and note your settings so you know what made the difference. PixVerse also suggests generating several takes from one image and keeping the best.

V6 can produce wide, medium, close-up and transition shots within a single generation. To keep a person looking the same across those shots, repeat the same identity details in every shot description; PixVerse's own example is "blonde hair and a blue silk shirt".

Templates, effects and mini-apps

If you would rather not write a prompt from scratch, PixVerse lists Templates, described as "ready-made creative starts", alongside a Character feature for consistent characters and a Lip Sync tool. Its site also names mini-apps including Sound Effect Generator, Magic Extend, Video Expand, Director Shots and Product Video Remix, although the pages we read do not explain them.

Effects follow the same two routes. PixVerse's platform documentation calls effects its signature feature: with an image, the effect is applied to your picture; in text-to-video, it is applied from the prompt. That documentation, written for the API platform, also says effects use a limited number of activations, that higher memberships get a larger quota and that free users do not get resets, so check the app before planning around a particular effect.

Credits, the watermark and which plan to pick

Every generation spends credits. PixVerse's subscription page does not state a cost per generation, and the V6 review advises checking the in-app estimate before scaling up. For a sense of scale, the review quotes PixVerse's platform (API) pricing at 18 credits per second at 1080p without audio and 23 with audio, so a 15-second 1080p clip with sound is 345 credits; it does not say the web app charges the same.

The free Basic plan costs nothing and gives 60 initial credits plus 30 daily credits, up to 540p and 2 generations at a time. Videos made on it carry a watermark. Daily credits renew at 0:00 UTC and expire at the end of the day.

Paid plans remove the watermark. Standard is $10 a month, or $8 a month billed yearly ($96), with 1,200 credits every 30 days plus 60 daily credits, up to 720p. Pro is $30 a month or $24 billed yearly ($288), with 6,000 credits, up to 4K and an off-peak mode that saves 30% credits. Premium is $60 a month or $48 billed yearly ($576), with 15,000 credits. Membership credits refresh every 30 days and unused ones expire; bonus credits from credit packs do not. Standard suits watermark-free social clips at 720p; for V6's full 1080p, the plan table points to Pro or Premium.

PixVerse's V6 FAQ says commercial usage rights depend on your subscription tier, so check before using clips in ads. Refunds are not given for dissatisfaction with results, and credit packs are non-refundable. Kuponaut has no PixVerse code yet: we have a partner link and have asked for a reader code. PixVerse's own saving is yearly billing, shown as up to 40% off.

Label AI video when you post it

Most large platforms, including TikTok, YouTube, Instagram and Facebook, ask creators to disclose realistic AI-generated or altered content, and offer a label or setting for it when you upload.

As a rule of thumb, label any clip showing realistic people, places or events that did not happen, and especially anything with a real person's face or a cloned-sounding voice. A short note in the caption, such as "made with AI", helps viewers even where a label is optional. A paid plan removes the PixVerse watermark, but that is not a reason to present the clip as real footage.

PixVerse or Hailuo AI?

Hailuo AI, MiniMax's video generator and also a Kuponaut partner, works in a similar way: text, image and reference inputs, credits, and several models behind one prompt box, including Veo 3.1 and Sora 2. Its own Hailuo 2.3 model makes 6- or 10-second clips at 768p or 1080p, at 25 credits for 768p and 6 seconds, or 50 credits for 768p and 10 seconds or 1080p and 6 seconds. MiniMax says Hailuo 2.3 responds better to motion commands, and Hailuo 1.0-Director is offered for camera direction.

The Hailuo 2.3 page does not describe generated audio, although Hailuo lists native synchronised audio for other models it offers, such as MiniMax H3, Veo 3.1 and Sora 2. PixVerse's case is V6 with built-in sound and clips up to 15 seconds, plus its templates and effects. Credits are not the same unit on the two services, so the fairest test is the same prompt and photo on both free tiers.

Related code pages

Links

Listed on
Product HuntAI Just BetterAlternativeToF6SGitHubZenodoPostYourStartupLaunchBeamAITNT NewsCrunchbaseSaaSHubIndie Hackers