dubery-ugc-fidelity-gatekeeper

To validate UGC prompts for product fidelity, you run a gatekeeper process on the prompt JSON file to return a binary PASS or REJECT verdict. This prevents wasted ad pipeline credits on incorrect or distorted brand visuals.

A UGC prompt gatekeeper verifies text prompts against product specifications before generation. It outputs a binary PASS/REJECT verdict with no patching, providing deterministic gating and specific reject reasons to guard against brand visual distortion.

To run a fidelity check, provide the file path to your UGC prompt JSON, such as a .tmp/UGC-001_ugc_prompt.json file. The gatekeeper processes this input and returns a verdict JSON structure containing fields for verdict, checks, and reject_reasons.

No, UGC prompt validation does not automatically fix or patch rejected prompts. The gatekeeper outputs a strict binary PASS/REJECT verdict to ensure deterministic gating, leaving any necessary prompt corrections to the UGC pipeline operator.

You need a fidelity gatekeeper when generating UGC visuals via platforms like kie.ai to ensure generated sunglasses match the real product. It prevents wasted generation credits by catching prompts that would distort brand visuals before processing.

A key limitation is that validating UGC prompts requires a specific file path to a UGC prompt JSON file. Additionally, the gatekeeper only provides a binary verdict and does not patch or repair prompts that fail the fidelity checks.