Is there a text-to-video API with no content filter?

Yes. Api18.dev's text-to-video models accept your prompt without moderation and return a video URL. You control resolution, aspect ratio, duration and seed, and you can run it as one synchronous call.

What parameters does the text-to-video endpoint accept?

Typical parameters are model, prompt, resolution (e.g. 720p/1080p), aspect_ratio, duration in seconds, and an optional seed for reproducibility. The exact options per model are listed on each model's docs page.

Text-to-Video API with No Content Filters

A text-to-video API with no content filters turns a written prompt into a moving clip without a moderation layer deciding whether your scene is allowed. With Api18.dev you describe the shot, set a few parameters, and get a finished video URL back from one request.

How text-to-video works here#

You POST a model id and a prompt, plus optional controls for resolution, aspect ratio, duration and seed. The model generates a clip and the result lands at data[0].url. The prompt is not content-filtered.

curl -X POST "https://api18.dev/v1/generate?wait=true" \
  -H "Authorization: Bearer $API18_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "<video-model-id>",
    "prompt": "your prompt here",
    "resolution": "720p",
    "aspect_ratio": "16:9",
    "duration": 5
  }'

# → returns the finished job; the video URL is at .data[0].url

Replace <video-model-id> with a real text-to-video id from the models page, which lists each model’s exact parameters and price.

The parameters#

prompt — the scene. Describe subject, action, camera move, lighting and mood. Up to 2500 characters.
resolution — e.g. 720p or 1080p. Higher resolution costs more per second.
aspect_ratio — 16:9 for landscape, 9:16 for vertical/social.
duration — clip length in seconds; total cost scales with it.
seed — set for reproducibility; omit for a fresh result each run.

Writing a good video prompt#

Lead with the subject and the single most important action.
Add one clear camera instruction (push-in, pan, static, orbit).
Specify lighting and time of day for consistent mood.
Keep it focused — a tight prompt beats a wall of adjectives.

Sync vs. async#

Use /v1/generate?wait=true for one-call, no-polling generation. For long renders, drop ?wait=true and poll GET /v1/jobs/{id} — the full pattern is in the NSFW video generation guide.

Have a source frame or character to keep consistent? Use image-to-video instead — it animates a still you provide.

Pricing#

Text-to-video is billed per second of output in USD (rate depends on resolution and model), times duration. Failed jobs are refunded. Try it free with $10 of welcome credit in the playground, then top up in USDT on Solana or Tron.