Tor MCP Proxy
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
| TOR_SOCKS5 | No | Dirección del proxy de Tor. Tor Browser usa el puerto 9150; para usarlo, define TOR_SOCKS5=socks5://127.0.0.1:9150. | socks5://127.0.0.1:9050 |
| TOR_LOG_PATH | No | Log de auditoría en formato JSONL. ⚠️ Registra URLs y búsquedas. | |
| TOR_MAX_RETRIES | No | Cuántas veces reintentar con un circuito nuevo. | 3 |
| TOR_DOWNLOAD_DIR | No | El único directorio donde escribe download_file. | ~/Downloads/tor-mcp-proxy |
| TOR_REQUEST_TIMEOUT_MS | No | Tiempo máximo de espera por petición en milisegundos (máximo 120000). |
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
} |
Tools
Functions exposed to the LLM to take actions
| Name | Description |
|---|---|
| fetch_pageA | Fetch an HTTP(S) or .onion page through Tor and return its text (or raw HTML). Supports GET/POST/PUT/PATCH and sticky sessions (same circuit + cookies) via session_id. |
| check_exit_ipA | Return the public IP address that destinations see (the Tor exit node of a fresh circuit). |
| search_onionA | Search the web through DuckDuckGo's .onion service, entirely inside Tor. Returns titles, URLs and snippets; results may include both clearnet and .onion sites. |
| download_fileA | Download a document, image, archive or data file through Tor into the server's download directory (TOR_DOWNLOAD_DIR). output_path is relative to that directory; existing files are never overwritten. |
| extract_linksA | Extract links from HTML (no network access). By default only .onion links are returned; set filter_onion: false for all links. Pass base_url to resolve relative links. Results are deduplicated and grouped by host. |
| page_metadataA | Summarise HTML (no network access): title, description, headings (H1/H2/H3), link counts (internal/external/.onion), forms and meta tags. |
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 6 tools
Most tools target distinct actions: fetching pages, checking exit IP, searching onions, downloading files, extracting links, and summarizing metadata. The main overlap is between fetch_page and download_file, since both retrieve remote content, though download_file has a clear local-save distinction. Offline analysis tools extract_links and page_metadata are also related but produce different outputs.
Five of six tool names follow a clear verb_noun snake_case pattern: fetch_page, check_exit_ip, search_onion, download_file, extract_links. page_metadata deviates by using a noun phrase rather than a verb, but the overall convention remains readable and mostly consistent.
Six tools is well-scoped for a Tor proxy server, covering fetching, searching, downloading, link extraction, metadata extraction, and exit IP checking. Each tool has a clear place and there is no obvious redundancy or bloat.
The surface covers the core Tor browsing lifecycle: fetch pages, search onion services, download files, parse HTML offline, and verify exit IP. Minor gaps exist, such as no explicit session/cookie clearing or circuit rotation tool, but agents can largely work around these using existing capabilities.