Help Center

Creating with AI Video

Generate clips from text or a still image in AI Image Studio — Runway-style image-to-video and text-to-video with model, duration, and resolution controls.

Volnyn generates short AI video clips from a text prompt, an uploaded still, or a frame you just created in AI Image Studio. Video lives in the same studio as images (/user/ai-image) — switch the model chip from an image model to a video model, or open /user/ai-video?prompt=… (redirects with video mode pre-selected).

This guide follows the same structure as Runway's Creating with Gen-4 Video and Getting Started with Generative Video articles: draft a prompt, configure settings, generate, iterate.

What Volnyn video can do

  • Text to video — describe scene + motion; no image required (model-dependent)
  • Image to video — your still is the first frame; the prompt describes motion
  • Extend / chain — build longer sequences by using the last frame of one clip as input to the next
  • Library — saved MP4s at /user/ai-videos (also reachable from the studio rail)

Typical wait time: 30 seconds to 2 minutes depending on model and queue. The page polls until your clip is ready.

Step 1 — Start in video mode

  1. Open AI Image Studio (logged in).
  2. Open the Model chip and choose a video model, for example:
    • Veo 3 Fast — faster, good quality (Google Gemini API)
    • Veo 3 — highest quality Veo tier
    • Veo 3 Lite — lower credit tier
    • Seedance 2.5, Wan 3.0, LTX-2.5, Grok Imagine Video — fal-hosted alternatives
  3. Or bookmark /user/ai-video?prompt=Your+idea to land in video mode directly.

AI Studio in video mode with model and duration chips

Step 2 — Draft the prompt

Text to video

Describe what you see and how it moves:

“Aerial drone shot over misty pine forest at dawn, slow forward glide, sun rays breaking through trees.”

Image to video

When you attach a still (upload, Library pick, or Animate from an image result), the image already defines subject, color, and composition. Your prompt should focus on motion and camera:

“Slow push-in, subject turns head slightly, soft bokeh in background.”

Do not re-describe everything visible in the frame — that fights the model.

Image-to-video: motion prompt vs still frame

Prompt tips (Runway-style)

  • Use camera verbs: push in, pan left, orbit, handheld, static
  • Use temporal order: “first… then… finally…” for multi-beat motion
  • Keep clips short in language — you are generating 4–20 seconds, not a feature film

Step 3 — Configure settings

Before Generate, set the chips for your target platform:

Setting Typical values Notes
Aspect ratio 16:9, 9:16, 1:1, 4:3 9:16 for Reels/TikTok; 16:9 for YouTube
Duration 4–20 seconds Depends on model (Veo vs Seedance caps differ)
Resolution 480p, 720p, 1080p+ Higher resolution costs more credits/time
Reference image Optional upload or Library id Required for strongest image-to-video control

Credit cost shows implicitly when generation starts — failed jobs that never begin usually do not charge; check Usage & Billing if something looks wrong.

Step 4 — Generate and wait

Click Generate. The studio shows a progress state while the provider renders your clip.

Video generation in progress

Rules while a clip is running:

  • Up to 3 pending videos per account at once — wait for one to finish before starting many more
  • Do not refresh aggressively; the client polls /ai-video/status/{id} automatically
  • Clips are downloaded to Volnyn storage when ready (provider URLs expire quickly)

Step 5 — Review, download, iterate

When complete, the player shows your MP4 with actions to download, rename, or delete.

Finished video in the studio player

To improve a clip:

  1. Note the one problem (too fast, wrong camera, artifact on hands)
  2. Adjust only the prompt line that controls that (usually motion or camera)
  3. Regenerate — or extract the last frame and run image to video again for continuity

For longer stories, chain clips: export frame → new prompt → new clip → edit together in your NLE.

Models at a glance

Model family Best for
Veo 3.x General cinematic text/image-to-video
Seedance 2.5 ByteDance path; check duration/resolution caps
Wan 3.0 Alternative motion styles
LTX-2.5 Fast Quick drafts
Grok Imagine Video xAI video path via fal

Switch models from the same chip menu — no separate app install.

Troubleshooting

Issue What to try
“Already have 3 videos generating” Wait for a pending job to finish
402 insufficient credits Add credits or use a lighter model (Veo Lite)
Motion ignores the image Shorten prompt to only motion/camera
Wrong aspect Set ratio before generate; re-crop may be needed

See also Why can't I generate a video? and Video resolution FAQ.

Still need help?

Open Get help in your account menu, or submit a request from /user/support while signed in.

← Feature Contact support →