Information
- OpenAPI version:
3.1.0
Overview
FlixML API 0.1.0
Section titled “FlixML API 0.1.0”Agent-native API for driving ComfyUI image and video generation.
Agent guide: read SKILL.md or GET /api/guide for the full agent workflow.
Core surfaces:
- Studio — generate single images or clips (
/api/image/generate,/api/video/generate). - Projects — structured stories: projects → scenes → shots → images → videos → final render.
- Characters — reusable identities with base prompts, LoRAs, and voices (
/api/characters). - LoRA training — train and manage custom models: datasets, start/status, checkpoints, samples (
/api/lora-training/*).
Workflows and providers are discovered live (GET /api/workflows, GET /api/providers).
Agent identity: send Authorization: Bearer <key> (keys come from
scripts/manage_agent_keys.py). An admin key sees everything. Any other key sees only the
jobs, media and projects it created, and only the characters in its allowlist; it can also
be limited to certain workflows and capped on concurrent jobs. config.json
security.require_api_key (default false) decides whether a request with no key is rejected
or served unrestricted. A key that is sent but unknown or revoked is always rejected.
Authentication
Section titled “Authentication”bearerAuth
Section titled “bearerAuth”Agent API key (see scripts/manage_agent_keys.py). Required when config.json security.require_api_key is true. A non-admin key only sees what it owns.
Security scheme type: http