- Home
- Skills
- Image & Video
- gpt image workflow
gpt image workflow
Workflow skill for the gpt-image-mcp MCP server - text-to-image, image editing, style-reference creation and iteration, with cost, privacy and timeout safeguards. The MCP server and API key are set up separately by you.
Free
Works with the AI Agents you already use
gpt image workflow
Example session with this skill installed
Use image1.png as the subject and image2.png for the style. Change the background of image1 to a futuristic laboratory but keep the person exactly as they are. Quality high.
- Read your context and instructions
- Compiled the gpt image workflow
- Generated the image
I have processed the image edit request using edit_image. I used image1.png as the primary subject and image2.png to guide the artistic style, applying a futuristic laboratory background while preserving the original subject's pose.
Image Edit Result
| Attribute | Detail |
|---|---|
| Status | Successfully generated |
| Output Path | ~/outputs/generated_1739856123.png |
| Model | dall-e-3 |
| Dimensions | 1024x1024 |
| Format | PNG |
| Settings | Quality: high, Fidelity: high |
The original file image1.png remains unchanged. The new version with the requested background and style reference has been saved to your local output directory.
Next steps
- Open the file at
~/outputs/generated_1739856123.pngto verify the background replacement. - If the laboratory details need more or less complexity, we can iterate by calling
edit_imageon this new result withinput_fidelityadjusted. - If the style is perfect, I can use this output as a reference for future generations to maintain consistency.
image1.png
PNG · 1536×1024
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
Running out of stock imagery or struggling with manual image editing? This skill enables your agent to interact with the gpt-image-mcp server to handle text-to-image generation and complex image manipulation via local files.
What it does
- Generate images from text descriptions with controls for size, quality, and aspect ratio.
- Edit local images by providing a source file and a prompt to modify specific elements.
- Style-reference mapping by using one image's aesthetics to influence the creation or edit of another.
- Background manipulation to replace or remove backgrounds using alpha-channel masks.
- System diagnostics via endpoint checks to verify model availability and API connectivity.
How it works
- Tool discovery: The agent identifies
generate_image,edit_image, orcheck_endpointwithin your MCP environment. - Context gathering: You provide a description for a new image or local file paths for an edit.
- Execution: The agent calls the tool with specific parameters like
quality,size, orinput_fidelity. - Reporting: The agent returns the local path and URI of the newly created file along with metadata.
Frameworks & tools
Requires the gpt-image-mcp MCP server (Node.js 22+) and an OpenAI-compatible API key. It uses local file paths for input and output, supporting PNG, JPEG, and WebP formats.
Why this beats prompting it yourself
Writing the correct tool parameters and managing file paths for image-to-image tasks is error-prone. This skill manages the technical logic for multi-image references, masks, and timeout handling, preventing duplicate billing from failed retries.
Use cases
- Create branding assets or logos from a text description.
- Modify existing site assets, like changing a daylight photo to a night scene.
- Apply a consistent art style across a set of product photos using a reference image.
- Generate transparent UI icons by specifying background parameters.
Known limitations
Does not install the MCP server or provide API keys. It cannot process web URLs directly, requiring local file paths for all image inputs.
Requires gpt-image-mcp server and Node.js 22+. No web URL support for images. Single image output per call. Masking requires user-provided alpha-channel PNGs.
How to use
- 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
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 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
Free forever
No account required to browse
Trust & safety
Security scan
Security scan passed 7 days ago
- Free with an Agensi account
Requires