With this MCP fal.ai Image Server, you can:
Generate images from text prompts using the generate-image tool with any valid fal.ai model ID (default: fal-ai/recraft-v3).
Customize parameters: model, image_size (square_hd, square, portrait_4_3, portrait_16_9, landscape_4_3, landscape_16_9; default landscape_4_3), num_images (up to 5), guidance_scale (default 3.5), num_inference_steps (default 28), and enable_safety_checker (default true).
Save images locally to ~/Downloads/fal_ai by default or a custom directory via FAL_IMAGES_OUTPUT_DIR, with full file paths returned.
Integrate with MCP-compatible IDEs such as Cursor and Windsurf (configured via npx and FAL_KEY).
Receive robust error messages for common issues like missing API key, model errors, or missing npx.
Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type @ followed by the MCP server name and your instructions, e.g., "@MCP fal.ai Image Server generate a logo for a coffee shop using model fal-ai/kolors"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
ARCHIVED — This repository is no longer maintained. Issues and pull requests are closed.
Generate images from text prompts using fal.ai and the Model Context Protocol (MCP). Integrates with AI IDEs like Cursor and Windsurf.
Supports any valid fal.ai model and all major image parameters (size, inference steps, guidance scale, safety checker, batch of up to 5).
Saves images locally with accessible file paths.
Simple configuration and robust error handling.
Related MCP server: GIMP MCP
Requirements: Node.js 18+, fal.ai API key
Configure MCP:
{
"mcpServers": {
"fal-ai-image": {
"command": "npx",
"args": ["-y", "mcp-fal-ai-image"],
"env": { "FAL_KEY": "YOUR-FAL-AI-API-KEY" }
}
}
}Run: Use the generate-image tool from your IDE.
generate an image of a red apple
generate an image of a red apple using model fal-ai/kolors
generate 3 images of a glowing red apple in a futuristic city using model fal-ai/recraft-v3, square_hd, 40 inference steps, guidance scale 4.0, safety checker on
Supported parameters: prompt, model ID (any fal.ai model), number of images, image size, inference steps, guidance scale, safety checker.
Images are saved locally; file paths are shown in the response. For model IDs, see fal.ai/models.
Generated images are saved to ~/Downloads/fal_ai by default (XDG-aware on Linux). Set FAL_IMAGES_OUTPUT_DIR to use a custom folder — images are saved to <your-folder>/fal_ai. The full file path for each image is included in the tool's response.
FAL_KEY is not set: Set your fal.ai API key as shown above.
Model errors: Double-check the model ID at fal.ai/models, or refer to the fal.ai API docs.
npx not found: Install Node.js 18+ and npm.
Other errors: Consult your MCP client logs.
Your API key is used only locally to authenticate with fal.ai.
No user data is stored or transmitted except as required by the fal.ai API.
MIT License © 2025 Madhusudan Kulkarni
Generate an image from a text prompt using a selectable text-to-image model.
| Name | Required | Description | Default |
|---|---|---|---|
| prompt | Yes | Text prompt describing the image to generate | |
| model | No | ID of the text-to-image model to use (any valid fal.ai model ID) | fal-ai/recraft-v3 |
| image_size | No | Size of the generated image | landscape_4_3 |
| num_images | No | Number of images to generate | |
| num_inference_steps | No | Number of inference steps | |
| guidance_scale | No | Classifier Free Guidance scale | |
| enable_safety_checker | No | Enable the safety checker |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
No annotations are provided, so the description carries the full burden of behavioral disclosure. It mentions 'selectable text-to-image model' but fails to detail critical aspects like rate limits, authentication requirements, cost implications, or output format (e.g., image URL, base64). For a generative tool with zero annotation coverage, this is a significant gap in transparency.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
The description is a single, efficient sentence that front-loads the core functionality without unnecessary details. It avoids redundancy and wastes no words, making it easy for an agent to parse quickly.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
Given the tool's complexity (7 parameters, generative nature) and lack of annotations or output schema, the description is insufficient. It omits behavioral traits like safety, performance, or output handling, leaving the agent under-informed about how to effectively invoke and interpret results from this tool.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
Schema description coverage is 100%, meaning all parameters are well-documented in the schema itself. The description adds no additional parameter semantics beyond what the schema provides, such as explaining trade-offs between model choices or image size effects. This meets the baseline for high schema coverage but doesn't enhance understanding.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description clearly states the tool's purpose: 'Generate an image from a text prompt using a selectable text-to-image model.' It specifies the verb ('Generate'), resource ('image'), and mechanism ('text-to-image model'). However, with no sibling tools mentioned, it lacks explicit differentiation from alternatives, preventing a perfect score.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
The description provides no guidance on when to use this tool versus alternatives, prerequisites, or exclusions. It merely states what the tool does without context for its application, leaving the agent to infer usage scenarios independently.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
generate-imageScored across 1 tool
With only one tool, there is no possibility of confusion or overlap between tools. The tool's purpose is clearly defined and distinct by default.
A single tool inherently follows a consistent naming pattern. The tool name 'generate-image' uses a clear verb_noun format, which would be consistent if more tools were added.
One tool is too few for a server named 'MCP fal.ai Image Server', which suggests a broader image-related domain. A single generation tool lacks coverage for expected operations like editing, upscaling, or listing models.
The server is severely incomplete for an image generation domain. It only provides image generation, missing essential operations such as image editing, model selection beyond 'selectable', or metadata retrieval, which limits agent workflows.
Generate AI images, videos, music, SFX & speech in any AI assistant. Results appear inline in chat.
Create images and videos from prompts, with options for image mixing, reference images, and start/…
LLM chat, text tools, image generation, editing, batch image jobs, and asynchronous video generation
AI image + video generation for agents: --flag prompt DSL, async generate/poll, x402 pay-per-use.