Plugins

Plugin types

Six kinds are supported today (see PLUGIN_TYPE_MAP in the app repo):

• highlight_detector — propose highlight spans from transcript + duration + config.
• caption_generator — produce caption text from segments and plain text.
• style_interpreter — map a natural-language style brief into renderer parameters.
• posting_strategy — suggest a schedule over clips and linked social accounts.
• pre_render — adjust style_params before FFmpeg runs.
• post_render — inspect the rendered file path and return optional warnings or metadata.

Manifest

Each plugin directory contains plugin.json with name, type, version, entrypoint (module.ClassName), and optional schema_version, compatible_core (PEP 440 specifier against core semver), permissions, and signatures. Optional .deps/ (from pip install --target) is prepended to sys.path before import.

{
  "name": "my_plugin",
  "type": "highlight_detector",
  "version": "1.0.0",
  "schema_version": "1",
  "compatible_core": ">=1.0.0,<2.0.0",
  "description": "What this plugin does.",
  "entrypoint": "main.MyPluginClass",
  "permissions": []
}

HTTP API

Authenticated users can list plugins and read configs. Admin-only routes install from Git, rescans the plugin directory into the database, toggles enablement, patches config_json, and uninstalls. Shapes are defined on the /api/plugins router in the backend — cross-check the live contract in Swagger.

Sandboxing

When config_json.sandboxed is true, execution uses a subprocess with a trimmed environment and JSON context passed in; the child prints JSON on stdout. Timeouts apply so runaway plugins cannot block the worker indefinitely.

Examples

Reference implementations ship under backend/plugins/ in Streamdat_app (for example keyword_highlighter and emoji_captions). Copy one as a template, change the manifest and class, then run POST /api/plugins/sync or install from a Git URL where your org enables that feature.