Change a product photo background with AI: check shape, color and labels

A reference-image workflow for product backgrounds: write the prompt, check shape and labels, and choose compositing when original pixels must stay intact.

Make a YouTube thumbnail with an AI background and editable text

Generate a background with room for a title, add editable text in an image editor, and check legibility before exporting your YouTube thumbnail.

Extract text into JSON, validate it, then export CSV

A Python workflow for LLM extraction with strict JSON validation, missing dates and leading-zero IDs. Includes downloadable helpers and offline tests.

Turn a landscape photo into a portrait layout: crop, pad or regenerate

Choose between cropping, adding space and AI reference-based generation for a portrait layout, with a worked sizing example and checks for subject distortion.

Turn a product photo into a short video: plan three separate shots

Build a three-shot product storyboard, use restrained image-to-video prompts, inspect changing product details, and assemble accepted clips in an external editor.

Translate a CSV with an LLM: preserve SKUs and check numbers

Translate description fields only, keep identifiers local, validate numbers and terminology, and resume from saved results. Python helpers and offline tests included.

Ofox Desktop setup: connect your AI tools and agents

Set up Ofox Desktop on an Apple Silicon Mac, find supported coding tools, connect your account and check a first request before adding more clients.

Turn One Webinar into a Blog, Email and Social Posts with AI

Repurpose a webinar with AI using an evidence brief, distinct channel angles, four worked examples, and checks that keep numbers and claims accurate.

Codex keeps testing but never finishes: what to change

Give Codex a clear stopping condition, inspect repeated test failures and preserve useful work. Compare completed tasks rather than token counts alone.

Use GPT Image 2.5 for UI mockups with consistent text

Create a UI image brief, preserve layout during edits and check exact copy with GPT Image 2.5. Separate a visual mockup from a working interface.