200 ChatGPT image and video shortcuts: my workflow

In my experiment with AI images and video, I worked with short requests I could return to: a profile portrait, a product image, an ad or a clip. This guide collects 200 shortcuts with a clear meaning for each, so you can choose a direction and add the details of your own project.

These are names for requests, rather than official ChatGPT commands. To make a shortcut such as /headshot consistent, define its meaning in the conversation. Image generation depends on the tools available in your account. Video needs a connected video tool, such as Higgsfield, and an eligible account.

How to use a shortcut without guessing its meaning

Choose a row, copy the shortcut and its description, then add a subject, reference and details. For example: /productshot: a studio image of the attached bottle, white background, square format. Preserve the label and packaging shape.

For repeated work, define a few shortcuts at the start of a conversation. You do not need all 200 at once:

In this conversation I will use personal shortcuts for image and video requests. They are not system commands. Use the meanings I define here:

/headshot = a professional portrait using a face photo I attach.
/productshot = a studio product image on a clean backdrop.
/staticad = a static ad using a message and text I supply.
/productreel = a vertical product video using my reference and instructions.

If I introduce another shortcut, I will send its meaning. Ask about a missing detail only if it changes the result. Preserve the person's identity, the product shape and supplied text. Do not invent features, prices, testimonials or results.

Use a generation tool available in this conversation. Before a credit-consuming action, state the tool and cost if it can be checked. If no suitable tool is available, say so and prepare a prompt for an external tool. Do not present a script or description as a generated file.

The shortcut does not replace a brief. For portraits, attach a clear face image. For products, supply a photo showing the label and proportions. Without a reference, describe what should appear and what must remain accurate.

OpenAI describes image creation through ordinary prompts or the image tool, where available, and editing uploaded images. ChatGPT image instructions.

What do you get: an image, mockup, diagram or video?

Most entries describe image directions. A mockup is a visualization. A 3D-looking illustration is not necessarily an editable model. /vectorflat describes an appearance; if you need SVG, print artwork or a 3D file, specify the format and check that the tool can deliver it.

Video requests include /ugc, /productspin and /animated in the opening group, plus 16 entries in the video section. They name the kind of clip you want. Duration, resolution, aspect ratio and audio support depend on the tool and model.

Connect Higgsfield to ChatGPT

Open Higgsfield's official connection page. If the plugin is available in your account, add it, sign in to Higgsfield and authorize the connection. Start a new conversation and explicitly request the tool. Official connection instructions.

The documentation states that an active subscription is required. Check eligibility and your balance before generating. Connected generation consumes credits; Unlimited access on the Higgsfield website does not make external-tool generation credit-free. Higgsfield's connected-tool credit explanation.

Connecting the plugin does not turn these 200 aliases into its commands. Describe the request, specify the tool and check the chosen model. This guide does not promise automatic selection of the best model or better results for every task.

If authorization fails, check the connection and whether the tool is available in the conversation. If images work but video does not, inspect the error, model availability, plan eligibility and credits. A general error does not establish that credits are exhausted.

A brief that preserves the details you need

I use a shortcut to return to a type of work, then specify the details that matter. For a portrait:

/headshot: Create a professional portrait from my attached photo. Preserve facial features, age, glasses and skin tone. Use soft window light, a simple light backdrop and a chest-up crop. Do not reshape the face. Provide a square image that also works in a circular crop.

For a product:

/productshot: Create a studio image of the attached product on white. Preserve its shape, proportions, colour and label. Do not add components or text. Use a square format with space around the product. If a label detail is unreadable in the reference, do not invent it.

For video, after connecting an eligible tool:

/productreel: Use Higgsfield for a vertical product video based on the attached image. First propose a model, duration, aspect ratio and credit estimate where available. Use a slow camera push-in and preserve the product. Do not add text, audio or product claims unless I supply them. If video generation is unavailable in this conversation, provide a prompt to copy and state that no clip was generated.

Three examples: a collectible figure, an exploded product and packaging

These shortcuts demonstrate different outputs you can request from the library. Each example includes a full prompt and an image generated with OpenAI's image tool for this guide. The products are fictional; these are not Higgsfield results or records of the original experiment.

/actionfigure: turn a reference person into a collectible figure

Use this for a personal content idea, a gift concept or character design. Attach a face photograph and change the output into a toy while asking to retain recognizable features.

/actionfigure: Turn the person in the attached photograph into a collectible action figure inside a box with a transparent window. Keep the bald head, black glasses, beard and light blue shirt recognizable as a sculpted toy. Show the full figure with dark trousers and white shoes. Add a miniature laptop and game controller in separate compartments. Navy packaging with lime accents, studio lighting, square format and the entire box visible. No text or logos.

Check recognizability, toy-like anatomy, accessories and the arrangement of packaging compartments. This output is a visual concept, not a production-ready toy model.

/productexplosion: show a product as separated components

This request creates a component illustration for a presentation or creative concept. Here, fictional headphones have one earcup's layers separated along a clear axis.

/productexplosion: Create a conceptual exploded image of fictional navy over-ear headphones with teal accents. Keep the headband and left earcup connected. On the right, space these layers clearly apart: ear cushion, outer shell, speaker disc, simple circuit board and inner housing. Preserve a recognizable headphone silhouette and orderly spacing. Light studio backdrop, crisp materials, square format. No labels, dimensions, text or technical specifications.

This is an artistic interpretation of a possible structure. For an actual product, provide a verified parts list or drawing and check the illustration against it. The generated image does not establish how real headphones are built.

/packagingdesign: develop a coordinated gift set

This example shows open and closed packaging, fitted compartments and coordinated contents. It helps explore a visual direction before producing manufacturing artwork.

/packagingdesign: Design a fictional coffee gift set. Use a rigid navy box, ivory interior and teal band with a simple leaf symbol without letters. In fitted compartments, place two ivory-and-teal coffee pouches and a dark espresso cup. Add a matching closed box behind it to show the exterior. Three-quarter product photograph on pale stone, square format. No brand name, text, price or existing logo.

Check colour consistency, compartment layout and space for the product information you need. This image does not include a dieline, dimensions or print file; prepare and verify those separately.

Copy a prompt as written or change its subject and colours. The opening alias shortens a request; it is not an official system command.

The 200-shortcut library

Each row describes a request direction. Copy the shortcut with its meaning, add details and inspect the output. Use Ctrl+F to find a shortcut on the page.

Official documents, commerce, property listings and technical drawings need additional checks. /passportphoto does not guarantee acceptance, /amazonready does not certify policy compliance, and a cutaway or dimensioned illustration is not verified engineering information. Use concepts as concepts, and preserve facts when presenting an actual product, property or service.

Six shortcuts to start with (6)

Portraits and profile images (18)

Product images (20)

Product concepts and 3D visuals (14)

Ads and marketing visuals (18)

Branding and logo concepts (16)

Social graphics (18)

Mockups and packaging (14)

Food, spaces and property concepts (14)

Illustration and art treatments (18)

Infographics and explanatory visuals (14)

Personal and playful concepts (14)

Video shortcuts: a connected generation tool is required (16)

Combine shortcuts into a workflow

Start with /productshot, choose a version, request /colorways, build a /staticad around the chosen colour and then move to /productreel. At each transition, attach or explicitly identify the approved image so the next request has a clear starting file.

Check the product, label and colour after every generation. A model can change a detail even when asked to preserve it. A small error in the first image may carry into the ad and video.

For a focused revision:

Continue from the last image I approved. Change only the background to light grey. Preserve the face or product shape, colours and text. Do not add objects. If you cannot preserve a detail exactly, state what changed.

Check before downloading or publishing

• Faces, hands, glasses and body proportions against the reference.
• Product shape, package contents, label and colours.
• Every word, price, figure and colour code in the image.
• File dimensions, aspect ratio and crops on the destination platform.
• In video: consistency between frames, camera movement, text and loop joins.

Use testimonials only when they are genuine and authorized. Generated UGC-style footage is not evidence of a real customer's experience. Label visualizations of properties, events or results so viewers understand what they are seeing.

Questions about ChatGPT image and video shortcuts

Can I just type /headshot in a new chat?

You can type it, but without a defined meaning there is no guarantee of consistent behavior. Include the definition or full brief and the reference image. These shortcuts shorten your own requests.

Do I need Higgsfield for images?

Not necessarily. Use ChatGPT's image tool where available. Higgsfield offers a connection to additional generation tools. Check availability and costs before starting.

Do all 200 shortcuts produce a finished file?

No. The output depends on the tool, supplied information and your review. A request may produce an image, a concept or a prompt. Text must not be presented as a generated image or video file.

Can I save the shortcuts for later?

Keep their definitions in a document and paste the relevant ones at the start of a task. If using persistent instructions, keep the rule brief rather than filling the space with an unused list. Check the limits in your account.

How do I keep the same person or product across outputs?

Attach the same reference, specify what to preserve and inspect each result. Changes in style, clothing or motion can still alter identity and proportions.

Can I create a logo or print artwork?

You can develop visual directions. Production needs checks of size, resolution, text, format and print requirements. A vector-looking image does not replace a verified vector file.

Start with one output

Choose a shortcut for the task in front of you, attach a reference and state what must remain intact. After the first result, change one detail at a time. Save the brief and approved version so you can return to them later.

For more requests to adapt to your workflow: NexAi's ChatGPT command library.