Implementation:Togethercomputer Together python Videos Create

Knowledge Sources	Together Python
Domains	Video_Generation, Generative_AI
Last Updated	2026-02-15 16:00 GMT

Overview

Concrete tool for generating videos from text prompts using AI models provided by the Together Python SDK.

Description

The Videos class provides API methods for creating AI-generated videos and retrieving job results. Video generation is asynchronous: create() submits a job and returns a job ID, while retrieve() polls for completion and returns the video URL. Supports configurable resolution, frame rate, duration, guidance scale, seed, keyframe images, and reference images. Both sync and async variants are provided.

Usage

Import this class when you need to generate videos from text descriptions or with image conditioning. The API uses a v2 endpoint and returns asynchronous job results.

Code Reference

Source Location

Repository: Together Python
File: src/together/resources/videos.py
Lines: 1-303

Signature

class Videos:
    def create(
        self,
        *,
        model: str,
        prompt: str | None = None,
        height: int | None = None,
        width: int | None = None,
        seconds: str | None = None,
        fps: int | None = None,
        steps: int | None = None,
        seed: int | None = None,
        guidance_scale: float | None = None,
        output_format: Literal["MP4", "WEBM"] | None = None,
        output_quality: int | None = None,
        negative_prompt: str | None = None,
        frame_images: List[Dict[str, Any]] | None = None,
        reference_images: List[str] | None = None,
    ) -> CreateVideoResponse: ...

    def retrieve(self, id: str) -> VideoJob: ...

Import

from together import Together

client = Together()
# Access via client.videos

I/O Contract

Inputs

Name	Type	Required	Description
model	str	Yes	Video generation model name
prompt	str	No	Text description of the desired video
height	int	No	Video height in pixels
width	int	No	Video width in pixels
seconds	str	No	Video duration (1-10 seconds)
fps	int	No	Frames per second (15-60, default 24)
steps	int	No	Denoising steps (10-50, default 20)
seed	int	No	Random seed for reproducibility
guidance_scale	float	No	Prompt adherence (6.0-10.0 recommended, default 8)
negative_prompt	str	No	What to avoid in generation
frame_images	List[Dict]	No	Keyframe images for guided generation
reference_images	List[str]	No	Style/composition reference images

Outputs

Name	Type	Description
create() returns	CreateVideoResponse	Contains job id for polling
retrieve() returns	VideoJob	Job status, video URL (when completed), cost

Usage Examples

import time
from together import Together

client = Together()

# Create a video generation job
response = client.videos.create(
    model="together/video-gen-model",
    prompt="A serene mountain landscape at sunset with flowing clouds",
    height=720,
    width=1280,
    seconds="5",
    fps=24,
    steps=30,
    guidance_scale=8.0,
    output_format="MP4",
)

print(f"Job ID: {response.id}")

# Poll for completion
while True:
    job = client.videos.retrieve(response.id)
    if job.status == "completed":
        print(f"Video URL: {job.outputs.video_url}")
        print(f"Cost: ${job.outputs.cost}")
        break
    elif job.status == "failed":
        print(f"Error: {job.error.message}")
        break
    time.sleep(5)

Related Pages

Page Connections

Double-click a node to navigate. Hold to expand connections.

Principle

Implementation

Heuristic

Environment