LangChain Integration

Use NextAPI as a custom tool in your LangChain agents. Wrap our REST API as a StructuredTool so any LangChain agent can generate videos as part of its execution chain.

Configuration

Set the following values in your LangChain configuration:

python

import os
import requests
import time
from langchain.tools import StructuredTool
from pydantic import BaseModel, Field

class VideoGenInput(BaseModel):
    prompt: str = Field(description="Text description of the video to generate")
    duration_seconds: int = Field(default=6, description="Video duration in seconds")
    resolution: str = Field(default="1080p", description="Output resolution")

def generate_video(prompt: str, duration_seconds: int = 6, resolution: str = "1080p") -> str:
    create_resp = requests.post(
        "https://api.nextapi.top/v1/videos",
        headers={
            "Authorization": f"Bearer {os.environ['NEXTAPI_KEY']}",
            "Content-Type": "application/json",
        },
        json={
            "model": "seedance-2.0-pro",
            "input": {
                "prompt": prompt,
                "duration_seconds": duration_seconds,
                "resolution": resolution,
            },
        },
    )
    create_resp.raise_for_status()
    video_id = create_resp.json()["id"]

    for _ in range(60):
        status_resp = requests.get(
            f"https://api.nextapi.top/v1/videos/{video_id}",
            headers={"Authorization": f"Bearer {os.environ['NEXTAPI_KEY']}"},
        )
        status_resp.raise_for_status()
        payload = status_resp.json()
        if payload["status"] == "succeeded":
            output = payload.get("output") or {}
            return output.get("url") or output.get("video_url") or video_id
        if payload["status"] == "failed":
            raise RuntimeError("NextAPI video generation failed")
        time.sleep(5)

    raise TimeoutError("NextAPI video generation did not finish within 5 minutes")

nextapi_tool = StructuredTool.from_function(
    func=generate_video,
    name="nextapi_video_gen",
    description="Generate an AI video from a text prompt using NextAPI",
    args_schema=VideoGenInput,
)

Setup Steps

1Install dependencies: pip install langchain requests pydantic
2Set your API key: export NEXTAPI_KEY=your-key-here
3Define the VideoGenInput schema and generate_video function as shown above.
4Create a StructuredTool wrapping the function.
5Add the tool to your LangChain agent's tool list.
6The tool creates the async task, polls by id, and returns a URL only after the task succeeds.

Test with cURL

Verify your setup by running this command with your real sk_live_ key:

Sample command — replace $NEXTAPI_KEY with your actual key from app.nextapi.top.

bash

CREATE_RESPONSE=$(curl -sS -X POST https://api.nextapi.top/v1/videos \
  -H "Authorization: Bearer $NEXTAPI_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "seedance-2.0-pro",
    "input": {
      "prompt": "a butterfly landing on a flower in slow motion",
      "duration_seconds": 6,
      "resolution": "1080p"
    }
  }')

VIDEO_ID=$(printf '%s' "$CREATE_RESPONSE" | jq -r '.id')

curl -sS "https://api.nextapi.top/v1/videos/$VIDEO_ID" \
  -H "Authorization: Bearer $NEXTAPI_KEY"

Keep NEXTAPI_KEY in server-side or local environment variables, not in shared notebooks. For strict setup details, see the third-party tools configuration guide.

Need help?

Contact our integrations team at support@nextapi.top or check the full API documentation.