More screenshots

    Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIMuseMuseDotsDots+21 more

    Nano Banana Prompting for Ad Frames

    1

    Write product-frame prompts with clear references, camera direction and usable copy space.

    $15

    Secure checkout via Stripe

    Money-back guarantee

    Converts to your local currency at checkout

    Security scanned

    Nano Banana Prompting for Ad Frames

    Example session with this skill installed

    Write a Nano Banana Pro prompt for a 9:16 hero frame: our cold-brew coffee can (label reads "NOMAD COLD BREW") on a sunlit stone ledge in Diriyah at golden hour. I have a packshot of the can as a reference. The Arabic headline goes top-right in post, so leave space.

    • Read your context and instructions
    • Compiled the nano banana prompting
    • Generated the image

    Summary

    Illustrative reference-driven prompt for Nano Banana Pro at 9:16. Check current resolution support in your provider. The can's identity comes from your packshot and the label text is quoted exactly. The top-right is kept empty for type set in post.

    Settings

    SettingValueWhy
    ModelNano Banana ProChosen for this example after checking current reference and text capabilities; confirm availability and price in your provider
    Aspect ratio9:16Vertical delivery
    ResolutionCheck live provider optionsSelect enough pixels for final delivery
    Referencepackshot.png (role: product)Identity comes from the image, not from a description

    Prompt

    Use the attached product photo as the exact can: same shape, colours and label.
    A single NOMAD cold-brew can stands upright on a weathered limestone ledge in the old mud-brick district of Diriyah, late golden hour. The can sits in the left third of the frame, slightly below centre. Behind it, soft out-of-focus mud-brick walls glow warm amber.
    Shot on an 85mm lens at eye level with the can, shallow depth of field: the label is tack-sharp and the background falls off into smooth bokeh. Low warm sun from the right rakes across the can, catching condensation droplets on the aluminium, with a soft natural shadow falling left.
    The label reads exactly "NOMAD COLD BREW" in its original typeface, crisp and legible.
    The upper-right area of the frame is open, calm sky and soft wall texture with nothing in it, leaving clear space for a headline.
    Photorealistic commercial product photography, natural colour, fine detail.
    

    Why it's written this way

    • Brief, not caption: subject, location, composition, lens, light and text are each stated once.
    • Positive framing: "the upper-right area is open, calm sky" works better than "no text, no objects top-right".

    Text rule: the can's label is product text, so it is quoted and generated. The Arabic headline is ad copy, so it's kept out and set in post.

    If it comes back wrong

    • Label garbled: re-roll once, then edit the image with only the label region changed.
    • The can drifts from the packshot: move the reference instruction to the first line and remove any descriptive colour words that conflict with it.

    packshot.png

    PNG · 1536×1024

    Generated

    Example file from a real run - the skill writes it into your workspace.

    Connects securely to your tools. The creator never sees your data.

    About this skill

    Use this skill after you know the product reference and visual direction. It turns those decisions into a Nano Banana prompt for a product frame, keyframe, reference sheet or end card. The prompt specifies what each reference controls, how the scene is lit and framed, which real product text must stay exact, and where editable ad copy will be added later.

    You receive a ready-to-send prompt and a check list for identity, packaging text, composition and reproducibility.

    Model availability, resolution and credit cost come from your current generation tool, not a fixed price table. Bring your own model access.

    Not for video prompts. On-screen ad copy must be handled in post. Logos are prone to distortion. Diegetic text requires manual verification against asset bibles.

    How to use

    1. 1

      Connect your agent to the Agensi MCP

      Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup

    2. 2

      Sign in to Agensi

      Your agent shows a sign-in link. Sign in and click Allow.

    3. 3

      Ask for what you need

      Your agent loads this skill and uses it on your task. It always asks before any payment.

    Reviews

    No reviews yet

    Be one of the first to try it. Every listed skill goes through the trust checks below.

    Security scanned

    Checked by our 8-point scan before listing

    Fresh listing

    Recently published to Agensi

    Money-back guarantee

    Doesn't do what its page promises? You get your money back.

    Trust & safety

    Security scan

    Security scan passed 1 month ago

    Permissions

    No special permissions declared or detected

    Listed1 month ago
    Updated1 month ago

    What's inside

    Frequently Asked Questions

    It gives the agent a repeatable prompt structure: subject, action, setting, composition, lighting, and exact product text. It helps diagnose text and material errors, but generated results still need review.

    The SKILL.md method is written for Nano Banana image generation. It includes Magnific model and tool notes, but the prompt principles also work in other agents that can send images and text to Nano Banana. Check your provider’s current model options before running.

    The ZIP contains SKILL.md, a commercial license, third-party notices, a cinematic-language reference, and a model-capabilities reference. It is a written prompting guide, not a model, API integration, or asset library.

    No, this skill is strictly optimized for high-fidelity still images; video generation requires a different prompt architecture not covered by this specific toolset.

    It instructs the agent to quote the exact string, specify its style and placement, then verify the result. Arabic joining and all product text still need human review. For critical logos or labels, composite the real mark in post if generation fails.

    Install the SKILL.md in a compatible agent and provide a product image or brief. Magnific library @name references are optional and only apply if you use that platform; a normal image reference also works.

    New in Image & Video