video/slideshow

Slideshow / Carousel pipeline — generate listicle content for TikTok/Instagram/LinkedIn.

Category: video
Source: workflows/video/slideshow.py

Input Schema

Field	Type	Default	Description
`film_grain`	`boolean`	`true`	Apply UGC-style film grain + warm tint for organic feel
`language`	`string`	`"en"`	Language code
`mixed_media`	`boolean`	`false`	Use video hook slide for Instagram mixed-media carousel
`niche`	`string`	`""`	Content niche
`num_items`	`integer`	`5`	Number of list items (slides)
`previous_topics`	`string[]`	—	Previously generated topics for dedup
`quality`	`string`	`"local"`	Quality preset (default ‘local’). Other values: ‘free’, ‘budget’, ‘standard’, ‘premium’. The ‘local’ preset uses in-process Diffusers / Kokoro / Whisper backends and falls through to remote per-stage when a backend isn’t installed, so machines without the local stack still produce a slideshow — they just call out to remote APIs for the missing pieces.
`regenerate`	`object`	—	When set, this run is a regeneration. Workflows may read `direction` / `keep` / `extra_instructions` to modulate prompts; the engine persists `parent_run_id` and `parent_variant_index` as run lineage columns.
`slide_duration`	`number`	`3.5`	Duration per slide in seconds
`slide_layout`	`string`	`"overlay_gradient"`	Slide layout style
`style_hint`	`string`	`""`	Visual style hint for image sourcing
`topic`	`string`	`""`	Listicle topic
`transition_duration`	`number`	`0.5`	Transition duration in seconds
`variants`	`integer`	`1`	Number of independent variant executions (1–10). When > 1, the engine runs the workflow N times with different sampling, producing N outputs.
`visual_style`	`string`	`""`	Visual style override

Output Schema

Field	Type	Default	Description
`hook_text`	`string`	`""`	Hook text
`hook_video_path`	`object`	—	Path to video hook slide (mixed-media)
`kind`	`object`	—	Variant card shape: video / carousel / image / text. Surfaced on the per-variant entry of the run-output API and used by gallery UIs to pick the right layout.
`pdf_path`	`object`	—	Path to multi-page PDF carousel
`platform_metadata`	`object`	—	Per-platform metadata
`slide_count`	`integer`	`0`	Number of slides
`slide_png_paths`	`string[]`	—	Paths to individual slide PNGs
`topic`	`string`	`""`	Generated topic
`video_path`	`object`	—	Path to stitched MP4 slideshow video

Task Pipeline

generate_listicle → source_slide_images → generate_slideshow_voiceover → generate_slideshow_bgm → generate_video_hook_slide → merge_slideshow_assets → compose_slides → export_slide_pngs → assemble_video_slideshow → apply_ugc_treatment → generate_pdf → generate_platform_metadata

Task	Description
`generate_listicle`	Generate listicle script with cross-run dedup.
`source_slide_images`	Source images for each slide based on image_prompts from script.
`generate_slideshow_voiceover`	Generate voiceover from the full listicle narration.
`generate_slideshow_bgm`	Generate background music.
`generate_video_hook_slide`	Generate a short video clip for the hook slide (Instagram mixed-media carousel).
`merge_slideshow_assets`	Merge parallel generation branches.
`compose_slides`	Apply text overlays to sourced images, producing final slide PNGs.
`export_slide_pngs`	Save individual slide PNGs as artifacts (carousel-ready).
`assemble_video_slideshow`	Stitch slides into MP4 with transitions, voiceover, and music.
`apply_ugc_treatment`	Apply film grain + warm tint to video and slide PNGs for organic UGC feel.
`generate_pdf`	Convert slide PNGs to a multi-page PDF for LinkedIn document carousels.
`generate_platform_metadata`	Generate per-platform metadata for the slideshow content.

Run-spec example

Save the YAML below as my-run.yaml, edit the values, and run with the CLI or POST it to the API. Required fields are uncommented; optional knobs are documented above the input: block — copy any line under input: and uncomment to set.

workflow: video/slideshow

# Optional fields — copy any line(s) under `input:` and uncomment to set:
# Apply UGC-style film grain + warm tint for organic feel
#   film_grain: true
#
# Language code
#   language: en
#
# Use video hook slide for Instagram mixed-media carousel
#   mixed_media: false
#
# Content niche
#   niche: ""
#
# Number of list items (slides)
#   num_items: 5
#
# Previously generated topics for dedup
#   previous_topics: []
#
# Quality preset (default 'local'). Other values: 'free', 'budget', 'standard', 'premium'. The 'local' preset uses in-process Diffusers / Kokoro / Whisper backends and falls through to remote per-stage when a backend isn't installed, so machines without the local stack still produce a slideshow — they just call out to remote APIs for the missing pieces.
#   quality: local
#
# Duration per slide in seconds
#   slide_duration: 3.5
#
# Slide layout style
#   slide_layout: overlay_gradient
#
# Visual style hint for image sourcing
#   style_hint: ""
#
# Listicle topic
#   topic: ""
#
# Transition duration in seconds
#   transition_duration: 0.5
#
# Visual style override
#   visual_style: ""
#

input: {}

Run it locally:

fab-workflow --from-file my-run.yaml

Or submit over the wire — the same file is the request body:

curl -X POST 'https://gofabric.dev/v1/workflows/runs?name=video/slideshow' \
  -H 'Authorization: Bearer fab_xxx' \
  -H 'content-type: application/yaml' \
  --data-binary @my-run.yaml

Every workflow also accepts the universal WorkflowInput fields — variants (1–10 fan-out) and regenerate (creative-direction hints with run lineage). See Run-specs (YAML / TOML / JSON) for the full top-level shape (metadata, priority, bundle, parent, etc.).

Warnings

Task merge_slideshow_assets has no Pydantic types — contract is opaque to consumers.