global/image-edit

Image editing workflow — modify images via img2img with natural-language prompts.

Category: global
Source: workflows/image/edit.py

Input Schema

Field	Type	Default	Description
aspect_ratio	string	""	Output aspect ratio override. Empty preserves the original.
image_path	string	required	Path to the source image (local file path or URL).
num_variants	integer	1	Number of edit variants to generate (1-4).
preserve	string[]	—	Aspects to preserve: ‘composition’, ‘colors’, ‘subject’, ‘style’, ‘lighting’.
prompt	string	required	Natural-language instruction describing the desired modification.
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.
strength	number	0.6	How much to deviate from the original image. 0.0 = nearly identical, 1.0 = almost entirely new. Default 0.6.
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.

Output Schema

Field	Type	Default	Description
edit_description	string	""	Description of what changed.
edited_image_path	string	required	Path to the primary edited image.
fidelity_score	number	0.0	How well the edit matched the prompt (0-10).
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.
original_description	string	""	Vision analysis description of the source image.
variant_paths	string[]	—	Paths to additional variants (if num_variants > 1).

Task Pipeline

analyze_source → apply_edit → evaluate_edit

Task	Description
analyze_source	Analyze the source image to understand composition, style, and content.
apply_edit	Apply the img2img edit to the source image.
evaluate_edit	Evaluate how well the edit matches the prompt.

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: global/image-edit

# Optional fields — copy any line(s) under `input:` and uncomment to set:
# Output aspect ratio override. Empty preserves the original.
#   aspect_ratio: ""
#
# Number of edit variants to generate (1-4).
# [min=1, max=4]
#   num_variants: 1
#
# Aspects to preserve: 'composition', 'colors', 'subject', 'style', 'lighting'.
#   preserve: []
#
# How much to deviate from the original image. 0.0 = nearly identical, 1.0 = almost entirely new. Default 0.6.
# [min=0.0, max=1.0]
#   strength: 0.6
#

input:
  # Path to the source image (local file path or URL).
  image_path: ""

  # Natural-language instruction describing the desired modification.
  prompt: ""

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=global/image-edit' \
  -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.).