1. Authenticate
Create an API key in your Medux account and send it as a Bearer token. Keep the key on your server, not in browser code.
Manage API keys →API Reference > Overview
Developer documentationMedux API Reference
Build AI video, avatar, voice, and media-processing workflows with Medux. Start with authentication and uploads, then use the endpoint guides to submit tasks and handle results.
Getting started with the Medux API
Use the same four-step pattern for most creation and processing workflows. Endpoint pages provide the exact fields and response contracts.
2. Upload input media
Request an upload URL, PUT the file to that URL, and use the returned file_id in a media endpoint.
Upload address API →3. Create a task
Choose a generation, cloning, or processing endpoint. Many media operations return an asynchronous task rather than a finished file.
Explore video generation →4. Read the result
Save the task ID, check its status, and use the output URL or produced asset ID when the task succeeds.
Task status API →A typical media request
Upload media only when an endpoint needs a file. Keep the returned file_id and task ID so your application can continue the workflow.
1. POST /api/app/v1/upload_urls
2. PUT {upload_url} # Upload the media bytes
3. POST /api/app/v1/generate/video/image_to_video
4. GET /api/app/v1/generate/tasks/{task_id}Send requests to https://api.medux.io with an Authorization: Bearer header. The creation endpoint, required fields, and task-status endpoint vary by capability; check the linked reference before implementing a specific flow.
Explore Medux API capabilities
Choose the category that matches the media you want to create or process.
AI video generation
Create videos from a text prompt or drive an avatar video with audio. Each endpoint describes its inputs, output, and task lifecycle.
Avatars and voices
Create reusable avatar and voice assets, then connect them to speech and video workflows.
Speech generation
Turn a script into audio with a voice asset and use the result in narration or lip-sync workflows.
Media processing
Process existing images and videos, including background and watermark removal, through task-based APIs.
Understand tasks, outputs, and errors
Media creation is often asynchronous. A successful task-creation response is the start of a job, not the finished output.
Store the returned data.task.id and poll the status endpoint for that task type. When the status becomes SUCCEEDED, read the output URL or produced asset ID described by that endpoint. Handle FAILED and EXPIRED as terminal states.
Medux responses include an application code and message. Check both the HTTP response and the application code before reading data. Review the error codes and the relevant task-status page for retry and failure handling.
REST API and MCP workflows
Use REST endpoints for direct product integrations. Medux also exposes media capabilities through MCP tools for compatible AI agents.
The endpoint pages document request fields, response fields, examples, and related operations. Where available, they also identify the corresponding MCP tool.