How to Use LTX 2.5 Fast Text-to-Video API: First Clip in 5 Minutes

TL;DR

• LTX 2.5 Fast Text to Video (ltx-2.5-fast/text-to-video) turns a text prompt into a 2–20 second video at 720p or 1080p, landscape or vertical, with native audio. Try it on the model page; new accounts get 200 free credits.
• Pricing is per second: 720p at $0.043/s (a 6-second clip is $0.258) and 1080p at $0.086/s; the generate_audio switch never changes the price.
• Everything runs on HiAPI /v1/tasks: submit for a taskId, poll, then take the video URL from output (temporary — save it).

What it does

• One sentence in, a sounding shot out: scene, motion, camera, and sound go into one prompt, and the audio generates with the picture.
• 2–20 seconds in 12 steps: duration takes 2, 3, 4, 5, 6, 8, 10, 12, 14, 16, 18, or 20 (default 6).
• Landscape or vertical, two resolutions: 16:9 / 9:16 at 720p / 1080p; fixed 25fps MP4.
• A sibling animates stills: ltx-2.5-fast/image-to-video takes a first frame (optionally a last frame) under the same pricing — see the image-to-video model page.

Here's a real clip (720p · 6 s, with native audio):

A barista pours steamed milk into a latte, creating a leaf pattern, close-up, soft morning light through a café window, gentle clinking cups and café ambience

Poster frame

ltx-2.5-fast/text-to-video, 720p, 6 s, 16:9, 1280×720, audio, $0.258. Watch the full clip →

Try this prompt →

Two ways to start

Browser, 3 steps: open the model page → paste the prompt, pick duration and resolution → generate. A 720p 6-second clip is $0.258.

API, 3 steps: POST /v1/tasks to submit → take the taskId → poll to success.

curl: your first clip

curl -X POST https://api.hiapi.ai/v1/tasks \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -H "Idempotency-Key: your-unique-key-001" \
  -d '{
  "model": "ltx-2.5-fast/text-to-video",
  "input": {
    "prompt": "A barista pours steamed milk into a latte, creating a leaf pattern, close-up, soft morning light through a café window, gentle clinking cups and café ambience",
    "resolution": "720p",
    "duration": 6,
    "aspect_ratio": "16:9",
    "generate_audio": true
  }
}'

This one bills 720p × 6 s = $0.258.

Python: submit, poll, download

import time, uuid, requests

BASE = "https://api.hiapi.ai/v1/tasks"
HEADERS = {
    "Authorization": "Bearer YOUR_API_KEY",
    "Content-Type": "application/json",
}

payload = {
    "model": "ltx-2.5-fast/text-to-video",
    "input": {
        "prompt": "A barista pours steamed milk into a latte, creating a leaf pattern, close-up, soft morning light through a café window, gentle clinking cups and café ambience",
        "resolution": "720p",
        "duration": 6,
        "aspect_ratio": "16:9",
        "generate_audio": True,
    },
}

resp = requests.post(
    BASE,
    headers={**HEADERS, "Idempotency-Key": str(uuid.uuid4())},
    json=payload,
    timeout=30,
)
resp.raise_for_status()
task_id = resp.json()["taskId"]

while True:
    task = requests.get(f"{BASE}/{task_id}", headers=HEADERS, timeout=30).json()
    print(task["status"])
    if task["status"] in ("success", "fail"):
        break
    time.sleep(5)

if task["status"] == "success":
    video = next(o for o in task["output"] if o.get("type") == "video")
    video_url = video["url"]   # temporary URL — download promptly
    mp4 = requests.get(video_url, timeout=120)
    with open("output.mp4", "wb") as f:
        f.write(mp4.content)
    print("saved output.mp4")

Parameters

Any field outside the table returns a 400.

Errors you'll actually hit

Terminal failures use CONTENT_POLICY_VIOLATION, TASK_FAILED, and TASK_TIMEOUT; failed tasks refund in full.

Before you ship

• Callbacks: pass "callback": {"url": "https://..."} at the request top level — video tasks run long, callbacks pay off.
• Persist outputs: the video URL in output is temporary; download immediately.
• Set duration and resolution explicitly: they drive the bill — don't rely on defaults.
• To animate stills: use the sibling image-to-video model page with a first frame for control.

FAQ

• Yes, audio comes with the video: generate_audio defaults on; describe the sound in the prompt (wind, clinking cups) and the track follows the picture.
• Yes, duration has 12 steps: enumerated 2–20; sending 7 or 9 returns a 400.
• Yes, price = per-second rate × seconds: 720p 6 s is $0.258 and 1080p 6 s is $0.516; muting doesn't save.
• Yes, $10 buys about 38 clips: at 720p, 6 seconds each.
• Yes, 9:16 gives vertical output: 720×1280 (720p) or 1080×1920 (1080p), ready for short-video feeds without cropping.
• Yes, start low and go high: iterate prompts and framing at 720p, then rerun the same prompt at 1080p for the final; the per-second price doubles but you only keep one final.
• Yes, for post-scoring, mute it: pass generate_audio: false when you'll add music or voice in post — clean footage at the same price.

Next: take the latte prompt and generate your first clip; four scene templates are in the text-to-video prompt guide.