awam-file
Exposes Docker volume metadata as a browsable awam://docker/ directory while the Docker engine is running, letting agents enumerate volumes through the unified URI namespace.
Exposes the local Hugging Face hub cache as awam://hf/cache/, so cached models and files can be listed, stat'ed and read through the same API as other resources.
Treats a Logseq knowledge base graph as a virtual backend (awam://knowledge/<graph>/), with a flat page view (_flat) and the ability to resolve knowledge pages to a local path for opening in external file managers.
Lists qBittorrent download tasks through its WebUI under awam://downloads/qb/, giving a virtual directory view of active torrents alongside the rest of the file namespace.
Exposes any configured rclone remote as a virtual backend (awam://rclone/<remote>/...) without code changes, and can materialize a remote file into a local download directory on demand.
This server does not currently have a deployable release on Glama. Follow the repository's setup instructions to install and configure it for your MCP client, including any required credentials and local dependencies.
awam-file
A unified resource gateway built around awam:// URIs. It exposes local disks, Everything search, rclone remotes, a Logseq knowledge base, qBittorrent downloads, and several other sources behind one consistent API — so file managers and AI agents can all talk to the same set of resources with the same semantics.
Clients that speak this API today:
Sigma File Manager (primary) — home-page entry plus
awam://...in the address barA File Manager Providers — dynamic backend discovery via
/api/backendsDirectory Opus (secondary) — jump to a resolved local path
AI agents — through the MCP stdio server
Features
One flat URI namespace covering local paths, UNC shares, search results, cloud remotes, knowledge pages, downloads and more.
Three surfaces over the same core: MCP (stdio), a localhost HTTP API, and a CLI.
Virtual backends (for example rclone) can materialize a file into a local download directory on demand.
Backend health and capability discovery at runtime — no client-side hardcoding of what exists.
Destructive remote operations are gated behind an explicit
allow_destructive=trueflag.
Related MCP server: Desktop Commander MCP
Requirements
Python 3.11 or newer
Optional external tools, only for the backends that use them: Everything (
es.exe), rclone, qBittorrent (WebUI), Prowlarr, Docker, WSL, kopia, Directory Opus. Backends degrade gracefully when their tool is unavailable.
Installation
pip install -e .Or run directly from a checkout with src on the import path:
export PYTHONPATH=src # Windows PowerShell: $env:PYTHONPATH = "src"
python -m awam_resource_gateway --helpUsage
MCP stdio server (for AI clients):
python -m awam_resource_gateway mcpLocalhost HTTP server (default 127.0.0.1:18787):
python -m awam_resource_gateway http
# UI: http://127.0.0.1:18787/filesystemsCLI examples:
python -m awam_resource_gateway backends
python -m awam_resource_gateway list "awam://knowledge/"
python -m awam_resource_gateway search "readme" --backend search
python -m awam_resource_gateway open-dopus "awam://knowledge/logseq-main"URI overview
URI | Purpose |
| Backend directory |
| Local filesystem path |
| UNC path ( |
| Everything search results as a virtual directory |
| Any configured rclone remote, without code changes |
| Knowledge base; |
| qBittorrent tasks (WebUI must be enabled) |
| Prowlarr search |
| Mapped shares plus |
| Recycle bin |
| Recently used files |
| User and Public download folders |
| Local Hugging Face hub cache |
| List distributions (stopped ones are not started) |
| Volume metadata while the engine is running |
| Snapshots when the CLI and a repo are available |
See docs/uri-spec.md for the full specification.
Configuration
Configuration lives in environment variables. A .env file at the project root is loaded at startup and is gitignored; variables already present in the environment take precedence over the file.
API keys and other secrets belong in .env (for example AWAM_PROWLARR_API_KEY) and must never be committed.
Executable paths for tool-backed backends can be overridden (in .env or the environment):
Tool | Variable | Default |
kopia |
|
|
Everything |
|
|
rclone |
|
|
Docker |
|
|
WSL |
|
|
PowerShell |
|
|
Directory Opus |
| platform default |
A bare command name is resolved through PATH; an absolute path must exist, otherwise the backend reports configured executable not found: ... instead of silently falling back.
MCP tools
res_backends, res_list, res_stat, res_read, res_search, res_resolve, res_open_in_dopus, res_push_to_downloads.
License
MIT. See LICENSE.
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
This server cannot be deployed
Maintenance
Related MCP Connectors
Reach your own computers from any AI app: files, search, commands, approvals and outage alerts.
Connect ChatGPT, Claude, Cursor or VS Code to your Windows PC: files, terminal, browser, apps.
Search and fetch curated AI agent skills, prompts and MCP servers.
Persistent memory, hybrid search and a goal graph for AI agents, over stdio or remote HTTP.
Related MCP Servers
- AlicenseNot gradedqualityDmaintenanceEnables AI assistants to search local files on Windows using the Everything search engine, supporting basic and complex file searches with filters like date, size, media type, and document type.4MIT
- AlicenseNot gradedqualityDmaintenanceEnables AI clients to manage files and documents, search code, run shell commands, and control processes locally across Windows, Linux, and macOS.1MIT
- AlicenseNot gradedqualityCmaintenanceEnables AI agents to remotely search, update, manage files, and run terminal commands on a desktop environment via Streamable HTTP with OAuth 2.0, supporting ChatGPT and web agent integrations.1,287,096 npm1MIT
- AlicenseNot gradedqualityBmaintenanceEnables AI agents to search, update, manage files, and run terminal commands on a computer, with remote access via Streamable HTTP and OAuth 2.0 for integration with ChatGPT and web agents.1,287,096 npm1MIT