PixelForge MCP
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
| GOOGLE_API_KEY | Yes | Your Google API key for Gemini and Imagen APIs. | |
| GOOGLE_CLOUD_PROJECT | No | Your Google Cloud project ID for Vertex AI features (optional). |
Instructions
Guidance the server publishes about itself, which clients place ahead of the tool catalog so the model reads it before choosing anything.
This server publishes no instructions, or was last inspected before Glama recorded them.
Capabilities
Features and capabilities supported by this server
Protocol revision2025-11-25
| Capability | Details |
|---|---|
| tools | {
"listChanged": true
} |
| logging | {} |
| prompts | {
"listChanged": true
} |
| resources | {
"subscribe": false,
"listChanged": true
} |
Tools
Functions exposed to the LLM to take actions
| Name | Description |
|---|---|
| generate_imageC | Generate an image from a text prompt using Google Gemini. |
| edit_imageC | Edit an existing image using a text prompt. |
| remove_backgroundA | Remove the background from an image. Uses AI to isolate the main subject and remove the background. Best results with clear subjects against distinct backgrounds. |
| transform_imageC | Transform an image using Pillow operations. |
| upscale_imageA | Upscale an image to a higher 1K/2K/4K size tier with a Gemini image model. The source is re-rendered at a higher size tier with an instruction to reproduce it exactly. Fine details can differ from the source; for exact pixel scaling use transform_image(operation="resize"). |
| advanced_editA | Edit an image with a specialized mode using a Gemini image model. Each mode wraps the prompt in an instruction tuned for that edit. An optional mask (white = region to change) is passed to the model as guidance; the boundary is interpreted semantically, not pixel-exactly. |
| analyze_imageC | Analyze an image and get a detailed description. |
| extract_textB | Extract text from an image using OCR. Uses Gemini's vision capabilities for high-quality text extraction. Works with screenshots, documents, signs, handwriting, etc. |
| detect_objectsA | Detect objects in an image with bounding boxes. Uses Gemini's zero-shot object detection. Returns bounding box coordinates normalized to 0-1000 scale as [y_min, x_min, y_max, x_max]. |
| compare_imagesB | Compare two or more images and analyze differences. Useful for A/B testing, design review, before/after comparison, and change detection. |
| optimize_promptA | Enhance a basic prompt for dramatically better image results. Uses AI to add details about lighting, composition, colors, textures, and artistic techniques. The #1 way to improve generation quality. |
| list_templatesB | List available prompt templates for image generation. |
| apply_templateC | Apply a prompt template with a subject to generate a ready-to-use prompt. |
| estimate_costA | Estimate the cost of an image operation. Shows pricing per model per operation to help with budgeting. |
| list_historyC | List recent generation history. |
| get_generation_detailsC | Get details of a specific generation by ID. |
| batch_generateA | Generate multiple images in parallel from a list of prompts. |
| list_available_modelsA | List all available image generation models with capabilities. Returns model details including speed, quality, resolution support, and best use cases. Use this to pick the right model for your task. |
| get_server_infoA | Get MCP server configuration, capabilities, and available tools. |
Prompts
Interactive templates invoked by user choice
| Name | Description |
|---|---|
No prompts | |
Resources
Contextual data attached and managed by the client
| Name | Description |
|---|---|
No resources | |
TDQS
Scored across 19 tools
Most tools have clearly distinct purposes: generate_image vs batch_generate, analyze_image vs detect_objects vs extract_text, and transform_image vs upscale_image are well differentiated. Some mild overlap remains between edit_image and advanced_edit, and between list_available_models and get_server_info, though descriptions provide enough guidance.
All tool names use snake_case, which is consistent throughout. Most follow a verb_noun pattern, but a few are adjective/noun-first or otherwise nonstandard, such as advanced_edit and batch_generate, keeping it from being perfectly predictable.
Nineteen tools is slightly above the ideal range, but the server covers a broad image platform spanning generation, editing, analysis, utilities, prompt assistance, cost estimation, and history. The count is defensible overall, though the surface is on the heavy side.
The server covers the core image lifecycle well: generation, batching, editing, upscaling, Pillow transforms, background removal, analysis, OCR, object detection, comparison, prompt templates, cost estimation, and history. Minor gaps exist around template/history management, such as no create/delete template or delete history operations.