AI Video Generator: Veo 3, Kling, Hailuo H3, Seedance, Wan

AI video generator: text to video and image to video with Veo 3 / 3.1 / 3.1 Lite, Kling 2.5–3.0, Hailuo 2.3 / H3, Seedance, Wan 2.2–2.6, LTX-2 (up to 4K), PixVerse v5 and Runway Gen-4 Turbo, billed per second of delivered video.

• 🎬 Text to video and 🖼️ image to video (first frame) with the same input.
• 🔊 Native audio on Veo 3, Kling 2.6 and Seedance 2.0.
• 💸 From $0.075 per second (Seedance 1 Pro Fast 480p, Wan 2.2 480p).
• ✨ Prompt optimization ($0.015): pick a use case (product, ads, social, UGC, story, cinematic, portrait) and get a complete, model-specific prompt.
• 🔑 No API key: no OpenAI, Google or ByteDance account needed.
• 🆓 Works with a free Apify account too: 1 generation per month (one 2 s video) with Seedance 1 Pro Fast, 480p, 2 s, billed at the normal price ($0.15). Every model and option on paid Apify plans.
• 🚫 Sexually explicit content is not allowed. Such prompts stop at once, at no cost.
• 🎯 Quality score for automation (optional, $0.02 per video): an AI checks every video for weird or broken details and for its match with your prompt, so you can publish without checking by hand.
• 💸 Pay only for delivered seconds of video. Failures are never charged.

Works from the Apify Console, the API, the JavaScript / Python clients, n8n, Make, Zapier and AI agents / MCP.

🎯 Quality score: run stories, ads and UGC on autopilot

Turn on qualityScore and every delivered video is checked by a multimodal AI before it reaches you, in two steps:

1. Coherence: nothing weird or broken (hands and fingers, faces, objects merging or appearing from nowhere, garbled text, illogical actions, morphing or flicker between frames).
2. Prompt match: the subject, product, setting, action and any quoted text of your original prompt are there (not the optimized prompt), and nothing was added that you did not ask for (text, logo, object, product color).

Each issue is graded like an art director looking at the image at normal size: critical (unusable: wrong product, something appearing from nowhere), major (visible at first glance) or minor (only seen when zooming in, such as watch hands or buckles).

Each row then has qualityScore (0-100), coherenceScore, promptMatchScore, issues (short list in English) and verdict:

Verdict	Rule	What to do
pass	no critical or major issue, quality >= 75	publish automatically
review	a major issue (first-glance defect, unrequested text...)	a quick human look
fail	a critical issue or quality < 50	discard or regenerate

No more checking every video by hand: in n8n, Make, Zapier or your own code, publish the pass rows of your daily stories, ads or product clips and regenerate the fail ones. $0.02 per scored video; if the checker is unavailable the video is delivered without a score and nothing extra is charged.

💡 Who is it for?

• Social media: TikTok, Reels, Shorts
• Ads and e-commerce: product videos from a packshot
• Developers and AI agents: one video API for every model

⚙️ How it works

1. You send a prompt, optionally a start image for image-to-video, and pick a model.
2. The Actor checks your maximum cost per run covers the full video, then generates it (from ~30 s to a few minutes).
3. Each video is copied to your Apify key-value store, so the link does not expire after an hour.
4. One dataset row per video: URL, model, options, duration.

💰 Pricing

Pay per event, per second of delivered video, no start fee. Paid Apify plans: every model below. Works with a free Apify account too: 1 generation per month (one 2 s video) with Seedance 1 Pro Fast, 480p, 2 s, billed at the normal price ($0.15).

Model	Option	Event	Price per second	Example
Wan 2.2 Fast (text-to-video or image-to-video, 5 s)	480p	video-lite-second	$0.075	$0.38 for 5 s
Wan 2.2 Fast (text-to-video or image-to-video, 5 s)	720p	video-standard-second	$0.20	$1.00 for 5 s
Seedance 1 Pro Fast	480p	video-lite-second	$0.075	$0.38 for 5 s
Seedance 1 Pro Fast	720p	video-standard-second	$0.20	$1.00 for 5 s
Seedance 1 Pro Fast	1080p	video-pro-second	$0.40	$2.00 for 5 s
Seedance 1 Lite	480p	video-lite-second	$0.075	$0.38 for 5 s
Seedance 1 Lite	720p	video-standard-second	$0.20	$1.00 for 5 s
Seedance 1 Lite	1080p	video-pro-second	$0.40	$2.00 for 5 s
Seedance 1 Pro	480p	video-standard-second	$0.20	$1.00 for 5 s
Seedance 1 Pro	720p	video-pro-second	$0.40	$2.00 for 5 s
Seedance 1 Pro	1080p	video-premium-second	$0.60	$3.00 for 5 s
Wan 2.5 (text-to-video or image-to-video)	480p	video-standard-second	$0.20	$1.00 for 5 s
Wan 2.5 (text-to-video or image-to-video)	720p	video-pro-second	$0.40	$2.00 for 5 s
Wan 2.5 (text-to-video or image-to-video)	1080p	video-premium-second	$0.60	$3.00 for 5 s
Kling 2.5 Turbo Pro (1080p)	—	video-pro-second	$0.40	$2.00 for 5 s
Kling 2.6 (with native audio)	—	video-premium-second	$0.60	$3.00 for 5 s
Hailuo H3 (MiniMax, 768p or 2K)	768p	video-pro-second	$0.40	$2.00 for 5 s
Hailuo H3 (MiniMax, 768p or 2K)	2k	video-premium-second	$0.60	$3.00 for 5 s
Veo 3 Fast (Google, with audio)	—	video-premium-second	$0.60	$2.40 for 4 s
Veo 3.1 Fast (Google, with audio)	—	video-premium-second	$0.60	$2.40 for 4 s
Seedance 2.0 (with audio)	480p	video-pro-second	$0.40	$2.00 for 5 s
Seedance 2.0 (with audio)	720p	video-ultra-second	$1.60	$8.00 for 5 s
Kling 3.0 Pro (1080p)	—	video-ultra-second	$1.60	$8.00 for 5 s
Veo 3 (Google, with audio)	—	video-ultra-second	$1.60	$6.40 for 4 s
Veo 3.1 (Google, with audio)	—	video-ultra-second	$1.60	$6.40 for 4 s
Wan 2.6 (text-to-video or image-to-video, up to 15 s)	720p	video-pro-second	$0.40	$2.00 for 5 s
Wan 2.6 (text-to-video or image-to-video, up to 15 s)	1080p	video-premium-second	$0.60	$3.00 for 5 s
LTX-2 Fast (Lightricks, with audio, up to 4K)	1080p	video-standard-second	$0.20	$1.20 for 6 s
LTX-2 Fast (Lightricks, with audio, up to 4K)	2k	video-pro-second	$0.40	$2.40 for 6 s
LTX-2 Fast (Lightricks, with audio, up to 4K)	4k	video-ultra-second	$1.60	$9.60 for 6 s
LTX-2 Pro (Lightricks, with audio, up to 4K)	1080p	video-pro-second	$0.40	$2.40 for 6 s
LTX-2 Pro (Lightricks, with audio, up to 4K)	2k	video-premium-second	$0.60	$3.60 for 6 s
LTX-2 Pro (Lightricks, with audio, up to 4K)	4k	video-ultra-second	$1.60	$9.60 for 6 s
Runway Gen-4 Turbo (image-to-video)	—	video-standard-second	$0.20	$1.00 for 5 s
PixVerse v5 (text-to-video or image-to-video)	360p	video-pro-second	$0.40	$2.00 for 5 s
PixVerse v5 (text-to-video or image-to-video)	540p	video-pro-second	$0.40	$2.00 for 5 s
PixVerse v5 (text-to-video or image-to-video)	720p	video-pro-second	$0.40	$2.00 for 5 s
PixVerse v5 (text-to-video or image-to-video)	1080p	video-ultra-second	$1.60	$8.00 for 5 s
Hailuo 2.3 (MiniMax, 768p or 1080p)	768p	video-pro-second	$0.40	$2.40 for 6 s
Hailuo 2.3 (MiniMax, 768p or 1080p)	1080p	video-pro-second	$0.40	$2.40 for 6 s

• A failed video is not charged. If nothing could be produced, nothing is charged.
• Prompt optimization (optional, optimizePrompt or a use case): $0.015 per optimized prompt (event prompt-optimization), charged once with the first delivered output that uses it. Not charged if the prompt writer is unavailable or nothing is delivered. Not available on the free Apify plan.
• Quality score (optional, qualityScore): $0.02 per scored video (event quality-score). Not charged if the checker is unavailable or the output failed. Not available on the free Apify plan.
• The billed duration is the delivered one (never more than requested). Before each video the Actor checks your remaining budget covers the whole requested duration, otherwise it stops without generating.
• Store discounts: on Apify Silver, Gold, Platinum and Diamond plans, some options cost up to 5 % less. The Pricing tab of the Store page shows the exact price for each plan.

📥 Input

Field	Type	Description
prompt	string	Describe the video: subject, action, camera movement, style, and sounds if audio is enabled.
model	string	Which AI model to use. Each option is billed at its own price, see the Pricing section of the README. Free Apify plan: 1 generation per month (one 2 s video) with Seedance 1 Pro Fast, 480p, 2 s, billed at the normal price ($0.15); other models and options are replaced by it.
imageUrl	string	Public URL of an image used as the first frame (image-to-video). Leave empty for text-to-video.
optimizePrompt	boolean	An AI prompt writer rewrites your idea into a detailed video prompt for the selected model (subject, composition, light, motion, camera, sound), keeping every detail you gave and any text in quotes. $0.015 per optimized prompt (event "prompt-optimization"), charged once with the first delivered output that uses it; nothing extra if the writer is unavailable. The prompt actually used is in the output ("prompt", with your text in "originalPrompt"). If the writer is unavailable, your own prompt is used.
preset	string	Style brief for the prompt optimization: product (e-commerce), ads, social post, UGC, story, cinematic, portrait. Picking a use case turns "Optimize my prompt" on.
qualityScore	boolean	A multimodal AI watches each video: coherence (hands, faces, objects, text, motion, illogical details) and match with your original prompt. Adds qualityScore, coherenceScore, promptMatchScore (0-100), issues and verdict (pass / review / fail) to each row, so you can publish only the "pass" ones automatically. $0.02 per scored video (event "quality-score"); nothing extra if the checker is unavailable.
quality	string	Shortcut for the resolution when "Resolution" is empty: draft = lowest, standard = model default, high = one step below the best, max = best the model offers. Higher quality costs more per second.
aspectRatio	string	Output aspect ratio. "auto" keeps the model default (for edits: the source image ratio). Not every model supports every ratio; an unsupported ratio stops the run before anything is charged.
resolution	string	Video resolution. Leave empty for the model default (the cheapest one). Higher resolutions cost more per second.
duration	integer	Length of the video. Supported values depend on the model (Wan 2.2 Fast: 5; Seedance 1 Pro Fast: 2 to 12; Seedance 1 Lite: 4 to 12; Seedance 1 Pro: 2 to 12; Wan 2.5: 5, 10; Kling 2.5 Turbo Pro: 5, 10; Kling 2.6: 5, 10; Hailuo H3: 4 to 15; Veo 3 Fast: 4, 6, 8; Veo 3.1 Fast: 4, 6, 8; Seedance 2.0: 4 to 15; Kling 3.0 Pro: 3 to 15; Veo 3: 4, 6, 8; Veo 3.1: 4, 6, 8; Wan 2.6: 5, 10, 15; LTX-2 Fast: 6, 8, 10, 12, 14, 16, 18, 20; LTX-2 Pro: 6, 8, 10; Runway Gen-4 Turbo: 5, 10; PixVerse v5: 5, 8; Hailuo 2.3: 6, 10). You pay per delivered second.
numberOfVariants	integer	How many separate videos to generate (1 to 4). Each successful one is billed; failures are free.
includeFailedVariants	boolean	Also write failed outputs (with the reason) to the dataset. Failed outputs are never billed.

Example: Cheapest text-to-video (Seedance 1 Pro Fast, 480p, 5 s)

{
  "prompt": "A paper boat drifting on a rainy street, cinematic",
  "model": "bytedance/seedance-1-pro-fast",
  "resolution": "480p",
  "duration": 5
}

Example: Image-to-video (Kling 2.5 Turbo Pro)

{
  "prompt": "The camera slowly orbits the product",
  "imageUrl": "https://example.com/product.jpg",
  "model": "kwaivgi/kling-v2.5-turbo-pro",
  "duration": 5
}

Example: Veo 3 Fast with audio

{
  "prompt": "A street musician plays violin at dusk, ambient city sound",
  "model": "google/veo-3-fast",
  "duration": 4,
  "resolution": "720p"
}

📤 Output

One dataset row per video:

{
  "model": "bytedance/seedance-1-pro-fast",
  "mode": "text-to-video",
  "prompt": "A paper boat drifting on a rainy street, cinematic",
  "resolution": "480p",
  "index": 1,
  "videoUrl": "https://api.apify.com/v2/key-value-stores/<STORE_ID>/records/video-1",
  "durationSeconds": 5,
  "contentType": "video/mp4",
  "chargedEvent": "video-lite-second",
  "chargedUnits": 5,
  "success": true
}

Export the dataset as JSON, CSV, Excel or HTML, or read it from the API.

🧠 Models

• Wan 2.2 Fast (text-to-video or image-to-video, 5 s) (wan-video/wan-2.2-fast): optional start frame image, resolutions 480p / 720p, 5 s, 2 aspect ratios.
• Seedance 1 Pro Fast (bytedance/seedance-1-pro-fast): optional start frame image, resolutions 480p / 720p / 1080p, 2 to 12 s, 7 aspect ratios.
• Seedance 1 Lite (bytedance/seedance-1-lite): optional start frame image, resolutions 480p / 720p / 1080p, 4 to 12 s, 7 aspect ratios.
• Seedance 1 Pro (bytedance/seedance-1-pro): optional start frame image, resolutions 480p / 720p / 1080p, 2 to 12 s, 7 aspect ratios.
• Wan 2.5 (text-to-video or image-to-video) (wan-video/wan-2.5): optional start frame image, resolutions 480p / 720p / 1080p, 5, 10 s, 2 aspect ratios.
• Kling 2.5 Turbo Pro (1080p) (kwaivgi/kling-v2.5-turbo-pro): optional start frame image, 5, 10 s, 3 aspect ratios.
• Kling 2.6 (with native audio) (kwaivgi/kling-v2.6): optional start frame image, 5, 10 s, 3 aspect ratios.
• Hailuo H3 (MiniMax, 768p or 2K) (minimax/h3): optional start frame image, resolutions 768p / 2k, 4 to 15 s, 6 aspect ratios.
• Veo 3 Fast (Google, with audio) (google/veo-3-fast): optional start frame image, resolutions 720p / 1080p, 4, 6, 8 s, 2 aspect ratios.
• Veo 3.1 Fast (Google, with audio) (google/veo-3.1-fast): optional start frame image, resolutions 720p / 1080p, 4, 6, 8 s, 2 aspect ratios.
• Seedance 2.0 (with audio) (bytedance/seedance-2.0): optional start frame image, resolutions 480p / 720p, 4 to 15 s, 7 aspect ratios.
• Kling 3.0 Pro (1080p) (kwaivgi/kling-v3-video): optional start frame image, 3 to 15 s, 3 aspect ratios.
• Veo 3 (Google, with audio) (google/veo-3): optional start frame image, resolutions 720p / 1080p, 4, 6, 8 s, 2 aspect ratios.
• Veo 3.1 (Google, with audio) (google/veo-3.1): optional start frame image, resolutions 720p / 1080p, 4, 6, 8 s, 2 aspect ratios.
• Wan 2.6 (text-to-video or image-to-video, up to 15 s) (wan-video/wan-2.6): optional start frame image, resolutions 720p / 1080p, 5, 10, 15 s, 2 aspect ratios.
• LTX-2 Fast (Lightricks, with audio, up to 4K) (lightricks/ltx-2-fast): optional start frame image, resolutions 1080p / 2k / 4k, 6, 8, 10, 12, 14, 16, 18, 20 s.
• LTX-2 Pro (Lightricks, with audio, up to 4K) (lightricks/ltx-2-pro): optional start frame image, resolutions 1080p / 2k / 4k, 6, 8, 10 s.
• Runway Gen-4 Turbo (image-to-video) (runwayml/gen4-turbo): needs a source image, 5, 10 s, 6 aspect ratios.
• PixVerse v5 (text-to-video or image-to-video) (pixverse/pixverse-v5): optional start frame image, resolutions 360p / 540p / 720p / 1080p, 5, 8 s, 3 aspect ratios.
• Hailuo 2.3 (MiniMax, 768p or 1080p) (minimax/hailuo-2.3): optional first frame image, resolutions 768p / 1080p, 6, 10 s.

🛟 Errors and reliability

• Invalid options (unsupported ratio, duration or resolution for the chosen model) stop the run before anything is generated or charged, with a message listing the supported values.
• If the AI provider is temporarily unavailable, the run stops right away (no time burned on the remaining outputs), nothing is charged, and the dataset contains one row with success: false, status: "service-unavailable" and a clear message. Retry a bit later.
• Runs whose maximum cost per run is too low for one output stop at once with an explanatory row (status: "budget-too-low"), at no cost.
• Free Apify plan: once the 1 generation of the month is used, or when the daily free-plan capacity is reached, runs stop at once with status: "free-plan-limit", at no cost. If free-plan runs are temporarily unavailable, the row says status: "paid-plan-required", also at no cost.
• If one output is refused (safety filter, unreadable source image), the others still run.

❓ FAQ

How long does a video take?

From ~30 s (Wan 2.2, Seedance Pro Fast) to several minutes (Kling 2.6, Kling 3.0, Veo 3). Set the run timeout to at least 30 minutes for long videos.

Do I need an API key (OpenAI, Google...)?

No. Your Apify account is enough: no key, no subscription, no quota to manage.

Can I use it on the free Apify plan?

Yes. Works with a free Apify account too: 1 generation per month (one 2 s video) with Seedance 1 Pro Fast, 480p, 2 s, billed at the normal price ($0.15), from your free monthly Apify credit. On a free plan only Seedance 1 Pro Fast, 480p, 2 s is used, one output per run: if you pick another model or option, the run uses it instead and says so in the output row (notice). Once the monthly generation is used (or when the daily free-plan capacity is reached), runs stop at once with status: "free-plan-limit", at no cost. Paid Apify plans get every model and option, without this limit.

Am I charged when a generation fails?

No. You pay only for videos that are delivered, per second of video. Invalid inputs are rejected before any generation starts.

How do I cap my spending?

Set Maximum cost per run in the run options. Before each video the Actor checks the remaining budget covers the full requested duration; if not, it stops without generating.

How long are files kept?

Outputs are copied to the run key-value store and follow your Apify data retention. Download them or copy them to your own storage for long-term use.

Can I call it from code, n8n, Make, Zapier or an AI agent?

Yes. Use the "Run Actor synchronously and get dataset items" API endpoint, the Apify clients, the integrations, or the Apify MCP server.

Who owns the generated content?

Use of generated content is subject to the terms of each model provider. Do not submit images you do not have the right to use.

Related actors

• Kling AI Video Generator: Only need Kling? Kling 2.5 Turbo, 2.6 with audio and 3.0.
• Seedance Video Generator: Only need Seedance? Cheap 480p to 1080p ByteDance Seedance videos.
• Nano Banana Image Editor: Create the first frame for image to video by editing a photo.
• AI Image Generator: Generate the starting image with Flux, GPT Image 2 or Seedream.

▶️ Start now

Pick a model, type a prompt, click Start.