Skip to main content
The /api/generate endpoint is the primary API for creating and modifying Remotion animation components. It supports two modes of operation:
  1. Initial Generation - Streaming response for creating new animations from scratch
  2. Follow-up Edits - Non-streaming response for making targeted edits or full replacements

​
Endpoint

POST /api/generate

​
Authentication

Requires OPENAI_API_KEY environment variable to be set. The API will return a 400 error if the key is missing.

​
Request Modes

​
Initial Generation Mode

When isFollowUp is false or not provided:
  • Uses streaming response for real-time code generation
  • Validates prompt is appropriate for motion graphics
  • Detects and applies relevant skills
  • Returns code via Server-Sent Events (SSE)
  • Prepends metadata event with detected skills

​
Follow-up Edit Mode

When isFollowUp is true and currentCode is provided:
  • Uses non-streaming response for faster edits
  • Returns either targeted edits or full code replacement
  • Supports error correction and self-healing
  • Preserves manual user edits when requested

​
Response Format

Initial Generation

Streaming response via SSE with metadata events

Follow-up Edits

JSON response with code, summary, and metadata

​
Common Use Cases

  • Generate a new animation from a text prompt
  • Edit existing animations (change colors, timing, text)
  • Fix compilation errors automatically
  • Provide visual context via frame images
  • Maintain conversation history for iterative edits

​
Error Handling

The endpoint returns specific error types:
  • validation - Prompt is not valid for motion graphics
  • edit_failed - Targeted edit could not be applied
  • Generic 500 errors for API/generation failures

​
Next Steps

Request Format

Learn about request parameters

Response Format

Understand response structure

Streaming

How streaming responses work

Build docs developers (and LLMs) love

Get started for freeTalk to us