Skip to main content
Glama

FAPIHUB MCP Server

An MCP server that gives AI agents (Claude Code, Claude Desktop, Cursor, VS Code, Windsurf and others) background-removal tools from FAPIHUB.

Ask your agent things like:

Remove the background from assets/product.jpg

Put every photo in ./team on a solid #F5F5F5 background

Blur the background of portrait.png and add a soft drop shadow to logo.png

The server reads images from your disk, sends them to the FAPIHUB API and saves the results next to the originals.

Preview

One photo, every tool:

Related MCP server: obrobka-mcp

Tools

Tool

What it does

Options

remove_background

Transparent PNG cutout of the subject

model: falcon (default), aurora, ghost

create_mask

Black-and-white subject mask

model

blur_background

Keeps the subject sharp and blurs the background

radius (1-50, default 10)

color_background

Replaces the background with a solid color

color

gradient_background

Replaces the background with a vertical gradient

top_color, bottom_color

add_shadow

Cutout with a drop shadow

shadow_color, offset_x, offset_y, blur_radius

get_usage

Your plan, its limits, and requests used and remaining

Every image tool takes image_path (JPG, PNG, WEBP, BMP or GIF, 1 KB to 10 MB) and an optional output_path. Without output_path, the result is saved as <name>_<operation>.<ext> beside the input and never overwrites an existing file.

Colors can be hex (#FF8C00, #FF8C0080) or RGB(A) (255, 140, 0 or 255, 140, 0, 128).

Each image tool call uses one request from your FAPIHUB plan. Invalid inputs (missing file, wrong format, bad size) are caught locally before anything is sent, so they don't use quota. get_usage is free, so agents can check how much quota is left before a large batch:

Plan: Pro (monthly, $15.00)
Status: active
Requests: 1,240 of 7,500 used (16.5%), 6,260 remaining
Successful: 1,225, failed (4xx, counted): 15
Rate limit: 60 requests per minute
Billing period: 2026-10-01 to 2026-11-01 (quota resets on 2026-11-01)

Setup

  1. Create an API key in your FAPIHUB dashboard.

  2. Install uv, which provides uvx.

  3. Add the server to your client using one of the configurations below.

Claude Code

claude mcp add fapihub -e FAPIHUB_API_KEY=bg_live_sk_... -- uvx --from git+https://github.com/Firdavs-coder/fapihub-mcp fapihub-mcp

Claude Desktop, Cursor, Windsurf and other JSON-configured clients

Add this to your client's MCP config (claude_desktop_config.json, .cursor/mcp.json, etc.):

{
  "mcpServers": {
    "fapihub": {
      "command": "uvx",
      "args": ["--from", "git+https://github.com/Firdavs-coder/fapihub-mcp", "fapihub-mcp"],
      "env": {
        "FAPIHUB_API_KEY": "bg_live_sk_..."
      }
    }
  }
}

VS Code

Add this to .vscode/mcp.json:

{
  "servers": {
    "fapihub": {
      "command": "uvx",
      "args": ["--from", "git+https://github.com/Firdavs-coder/fapihub-mcp", "fapihub-mcp"],
      "env": {
        "FAPIHUB_API_KEY": "bg_live_sk_..."
      }
    }
  }
}

Configuration

Environment variable

Required

Default

FAPIHUB_API_KEY

Yes

FAPIHUB_BASE_URL

No

https://fapihub.com

FAPIHUB_TIMEOUT

No

120 (seconds)

Development

git clone https://github.com/Firdavs-coder/fapihub-mcp
cd fapihub-mcp
uv sync
uv run pytest

To try the server against a local FAPIHUB instance:

FAPIHUB_API_KEY=bg_live_sk_... FAPIHUB_BASE_URL=http://127.0.0.1:8000 uv run fapihub-mcp

License

MIT

Available Tools

7 tools
add_shadowB

Cut out the subject and add a drop shadow behind it on a transparent background.

Inputs: image_path (required), offset_x, offset_y, blur_radius, +2 more. Returns: result

blur_backgroundA

Keep the subject sharp and blur the background (portrait-style bokeh).

Inputs: image_path (required), radius, output_path. Returns: result

color_backgroundB

Replace the background with a solid color.

Inputs: color (required), image_path (required), output_path. Returns: result

create_maskB

Create a black-and-white subject mask (white = subject, black = background) for an image.

Inputs: image_path (required), model, output_path. Returns: result

get_usageA
Read-onlyIdempotent

Show the FAPIHUB plan, its limits, and how many requests are used and remaining.

Free: does not use quota. Check it before processing many images.

Inputs: no named parameters. Returns: result

gradient_backgroundB

Replace the background with a vertical gradient from top_color to bottom_color.

Inputs: top_color (required), image_path (required), bottom_color (required), output_path. Returns: result

remove_backgroundB

Remove the background from an image, producing a transparent PNG cutout of the subject.

Inputs: image_path (required), model, output_path. Returns: result

Tool Schema Changelog

Recent tool additions, removals, and schema changes observed during successful MCP inspections.

Loading tool schema changelog…

    TDQS

    A3.7/5.0

    Scored across 7 tools

    Disambiguation5/5

    Each tool has a clearly distinct image-editing operation: background removal, mask creation, background blur, solid color replacement, gradient replacement, shadow addition, and usage inspection. While blur/color/gradient all manipulate backgrounds, their descriptions make the output differences explicit, and create_mask versus remove_background is also clear.

    Naming Consistency4/5

    Most tools follow a verb_noun pattern such as remove_background, create_mask, blur_background, color_background, and add_shadow. The one deviation is gradient_background, which reads more like noun_noun than verb_noun, but the overall convention remains readable and predictable.

    Tool Count5/5

    Seven tools is well-scoped for an image background processing API. Each tool corresponds to a meaningful operation, and there is no redundant filler.

    Completeness4/5

    The surface covers the core background-manipulation workflows: removal, masking, blur, solid color, gradient, shadow, plus usage inspection. A custom image background replacement or compositing tool would round it out further, but the existing set covers the primary domain operations.

    Maintenance

    ActivityMaintained
    ResponsivenessNo recent community issues

    Related MCP Connectors

    Related MCP Servers

    • A
      license
      Not graded
      quality
      B
      maintenance
      Enables AI agents to process images locally via file paths—converting, resizing, removing backgrounds, smart cropping, upscaling, reading or stripping metadata, and batch processing—without files ever leaving the device.
      MIT
    • A
      license
      A
      quality
      C
      maintenance
      Enables AI agents to perform local image operations such as circle cropping, cropping, resizing, compressing, converting, and inspecting images without uploading files or making network calls.
      7
      61 npm
      MIT
    • A
      license
      A
      quality
      B
      maintenance
      Local file tools for agents. Read and convert documents (DOCX, PDF, HWP, HWPX, EML), edit PDFs, resize and clean images, cut out backgrounds, make QR codes. Nothing is uploaded.
      6
      5
      MIT