# faktry [Health: Active]

**Category:** 💻 Developer Tools  
**Repository:** https://github.com/faktry/faktry-mcp  
**GitHub Stars:** 1  
**Views:** 0  
**Installs:** 0  
**Upvotes:** 0  
**Directory Page:** https://allmcps.com/mcp/faktry

## Description
Official faktry MCP server: generate and edit images, video, audio, and PDFs.

## Claude Desktop Quick Installation
Install path detected from listing signals. Uses `npx` (confidence: high):

```json
"mcpServers": {
  "faktry": {
    "command": "npx",
    "args": ["-y","@faktry/mcp"],
    "env": {
      "FAKTRY_API_KEY": ""
    }
  }
}
```

**Requires environment variables:** `FAKTRY_API_KEY` — the values above are empty placeholders; fill in real credentials before running (see the repository for what each one is for).

## Documentation

## What faktry MCP server does

The faktry MCP server exposes faktry’s media operations through the Model Context Protocol. It covers image, video, audio, PDF, and 3D workflows, allowing an MCP client to select tools based on a user’s request rather than requiring manual tool invocation.

Image capabilities include generation, editing, inpainting, outpainting, upscaling, background removal, effects, collages, watermarks, and smart cropping. Video tools support text-to-video and image-to-video generation, avatars, motion control, subtitles, reframing, stabilization, upscaling, and scene detection. Audio tools include speech, music, sound-effect, and voice-cloning generation, along with mixing, transcription, and trimming. PDF tools create, merge, split, compress, extract, and convert Markdown documents. Two 3D tools generate assets from text or images.

The faktry MCP server also includes utility operations for checking account credits, saving media to a library, listing saved assets, and tracking asynchronous jobs.

## How it works

The hosted endpoint is `https://faktry.ai/api/mcp`, so most clients connect without running a local process. Clients that only support stdio can use the `@faktry/mcp` npm bridge, which forwards requests to the hosted service and requires Node 18 or newer.

Authentication can use OAuth or a static API key. OAuth is supported by the documented one-click setup flows, while configuration-based clients send the key as a bearer token. Usage is charged against the connected faktry account’s credit balance.

Many generation operations are asynchronous, especially video and some image or audio tasks. The initial call returns a `job_id`; the client can pass that identifier to `faktry_job_status` until the job completes and provides a result URL.

## Setup and configuration

For remote-capable clients, add `https://faktry.ai/api/mcp` as an HTTP MCP server and complete the browser sign-in flow where OAuth is supported. API-key setups create a key at faktry’s API key page and send it in an `Authorization: Bearer ...` header.

The faktry MCP server can be configured directly in Claude Desktop, Cursor, Claude Code, VS Code, Windsurf, Continue, and Codex using the examples in the project documentation. Zed and other stdio-only clients can run:

```bash
FAKTRY_API_KEY=YOUR_API_KEY_HERE npx -y @faktry/mcp
```

The API key may be omitted from the bridge environment when using browser-based OAuth. Windsurf uses `serverUrl` rather than `url` in its configuration. Clients may need a restart or reload before newly added tools appear.

## Tools and capabilities

- `faktry_image_*`: image generation, editing, compositing, and enhancement.
- `faktry_video_*`: video generation and editing, including subtitles and reframing.
- `faktry_audio_*`: speech, voice, music, sound, transcription, and audio editing.
- `faktry_pdf_*`: PDF creation, conversion, and document manipulation.
- `faktry_3d_*`: text-to-3D and image-to-3D generation.
- `faktry_job_status`: checks an asynchronous job and returns its result URL when complete.
- `faktry_library_save` and `faktry_library_list`: persist and retrieve media across sessions.
- `faktry_credits`: reports credit balance and current-month usage.

## Limitations and notes

A faktry account and available credits are required for usage. The remote connection avoids local server management, but stdio-only clients need the npm bridge and Node 18 or newer. Long-running tasks are not returned as immediate media results; they require polling with the returned job ID.

_Full upstream README: https://allmcps.com/mcp/faktry/readme_