Search VisionStory documentation

No documentation matched “”.

Try a feature or resource name such as , , or .

VisionStoryDevelopers
Get API key
VisionStory AI video API

From script to speaking video, in one request.

Build talking-avatar video workflows without maintaining a media pipeline. Call the VisionStory REST API directly, use the Python SDK or CLI, or give the complete generation flow to an AI agent.

Create your first video Explore the API
Input
Text or audio
Output
MP4 video
Interfaces
API · SDK · CLI · MCP
POST/api/v1/video
from visionstory import VisionStoryClient, build_video_payload

client = VisionStoryClient.from_env()
video = client.generate_video(build_video_payload(
    avatar_id="your_avatar_id",
    text="Welcome to VisionStory.",
    voice_id="your_voice_id",
))
client.download(video["video_url"], "result.mp4")
✓Requestaccepted
✓Rendercompleted
▶result.mp4ready

Start with your workflow

Choose by what you want to do.

Building an application? Start with the Python SDK. Working in a terminal or CI job? Use the CLI. Want an AI agent to handle the workflow? Choose MCP or the Agent Skill.

sdk

Build in Python

Best for applications

Typed helpers for authentication, uploads, polling, and downloads with no runtime dependencies.

pip install visionstoryOpen guide
cli

Automate from the shell

Best for terminal & CI

Create videos, inspect tasks, and download results from scripts or CI without writing an app.

curl -fsSL https://developers.visionstory.ai/cli | bashOpen guide
mcp

Give tools to agents

Best for AI agents

Connect VisionStory through MCP or install the Skill so coding agents can generate media for you.

uvx visionstory-mcpOpen guide

Agent ready

Works where your agents work.

Copy one setup message for your coding agent, or connect the MCP server and expose VisionStory as a reusable tool.

VisionStory CLI

Install once. Use every API from your terminal.

The installer upgrades to the latest PyPI release in an isolated environment. Your existing Python projects stay untouched.

curl -fsSL https://developers.visionstory.ai/cli | bash
macOS · Linux · WSLInstallation options
Send this to Claude Code
Set up VisionStory so I can generate talking-avatar videos from here. Guide me as a first-time user and do not skip setup steps.1. Install the skill:   npx skills add visionstory-ai/skills --skill visionstory-api2. Help me configure authentication before making an API request:   a. If I do not have a key, direct me to https://www.visionstory.ai/openapi and pause while I create one.   b. Detect my OS and shell, then give me one exact command to set VISIONSTORY_API_KEY locally and persist it for future terminals.   c. Ask me to enter the key only in my own terminal. Never ask me to paste the key into chat, and never print, log, or commit it.   d. After I confirm, verify only that the variable is present and run a read-only credits check. If authentication fails, stop and help me fix it.3. Discover current models, avatars, and voices instead of guessing IDs.4. Use the visionstory-api skill to create and download the video. Do not delete anything unless I explicitly request and confirm it.

New here? Create an API key, then paste this into Claude Code. The agent will detect your terminal, show the exact setup command, verify access without revealing the key, and only then create the video.

One complete workflow

The shortest path from input to publish.

VisionStory handles the media pipeline between a script and a finished video. Your application stays focused on the experience around it.

  1. 01

    Choose an avatar

    Use a public avatar or create one from a portrait.

  2. 02

    Add the script

    Provide text or recorded audio, plus a voice.

  3. 03

    Generate

    Submit once; VisionStory renders asynchronously.

  4. 04

    Download

    Poll the task and save the finished MP4.

API capabilities

Everything behind the request.

Start with talking avatars, then reach for voices, image generation, frontier video models, and reusable media assets as your workflow grows.

4 endpoints

Talking Avatar Video

Generate a lip-synced talking-avatar video from a text script or recorded audio.

POST/api/v1/videoGET/api/v1/videoGET/api/v1/videosDELETE/api/v1/video
Open guide
3 endpoints

Avatars

List public AI avatars or create a reusable custom avatar from one portrait.

Open guide
3 endpoints

Voices

List public voices or clone a custom AI voice from an audio sample.

Open guide
1 endpoint

Models

Discover supported talking-avatar models, resolutions, and aspect ratios.

View reference
1 endpoint

Billing

Read the current subscription plan and remaining generation credits.

View reference
6 endpoints

AI Video

Generate text-to-video, image-to-video, and reference-guided AI clips.

Open guide
1 endpoint

Text to Speech

Convert text into natural MP3 speech with any public or cloned voice.

Open guide
2 endpoints

Image Generation

Generate or edit AI images from text prompts and optional reference media.

Open guide
3 endpoints

Assets

Upload and reuse images, audio, and video across generation requests.

View reference
Need the machine-readable surface?Browse every endpoint

Your first successful call

Make the first request.

Start with an API key, run the ready-made example, and use the returned video_id to follow your render.

Open quickstart Get API key
Request path Ready to run
  1. 1
    AuthenticateVISIONSTORY_API_KEY
  2. 2
    Send the request

    POST/api/v1/video

  3. ✓
    Receive a task IDvideo_123456
200 successJSON response
{
  "data": {
    "video_id": "video_123456"
  },
  "message": "success",
  "server_time": "2026-08-24T08:00:00Z"
}