Skip the install. Get this working in under 2 minutes.
Start a free trial on cloud.anythingmcp.com, add the Hetzner Storage Box (File Search) in one click, then point your AI client (Claude, ChatGPT, Copilot or Cursor) at the generated MCP endpoint. No Docker, no git clone, zero engineering experience required.
AnythingMCP is listed in Claude's connector directory: add it to Claude in one click (opens in a new tab), then sign in and choose which MCP servers Claude may use.
Summary
Let Claude search inside the files of a Hetzner Storage Box over SFTP on port 23, with a sub-account limited to one folder: full-text index of PDF, Excel, Word and Outlook files, synced by AnythingMCP File Search.
Try asking
Example prompts for Hetzner Storage Box (File Search)
Click any prompt to copy it. Paste into Claude, ChatGPT, Cursor, Gemini, Copilot or OpenClaw to run it against this connector.
Claude is AI and can make mistakes. Please double-check responses.
Hetzner Storage Box in Claude
Short answer: add the Storage Box to AnythingMCP as a File Search connector over SFTP on port 23, ideally with a sub-account limited to one folder. AnythingMCP indexes the full text of that folder and keeps the index in sync. Then connect AnythingMCP to claude.ai, the Claude desktop app and Claude Code. Claude can find files by name and by content, read the pages it needs and give you the file itself, attached or as a temporary download link. Nothing is ever written to the Storage Box.
An AI assistant cannot open a Storage Box on its own, and one that could would have to download every file to answer one question. The index is what makes a question like "what does article NT-601 cost in the current price list?" take seconds.
What you need
- The host name of the Storage Box:
uXXXXXX.your-storagebox.de. - SSH support enabled in the Storage Box settings. SFTP runs on port 23, not 22.
- The main user, or better a sub-account limited to one folder (Robot → Storage Box → Sub-accounts).
- An AnythingMCP workspace: AnythingMCP Cloud (7 days free) or your own self-hosted install.
A sub-account sees only its folder. Create one for AnythingMCP and the connector cannot read the rest of the Storage Box, whatever you tick in the wizard.
Set it up
Step 1 — Set up the File Search connector
In AnythingMCP, open the Marketplace and pick Hetzner Storage Box (File Search): it opens the File Search wizard with SFTP and the settings for the Storage Box. (Or open Connectors → New → File Search and use the Hetzner Storage Box shortcut.)
- Connect. Enter
uXXXXXX.your-storagebox.de, port 23, the sub-account and its password, and the folder to use. - Test. AnythingMCP connects, signs in and lists the folder before anything is saved, and shows a checklist (address, reachability, encryption, server identity, login, folder) with what to fix when a step fails. It also shows the server's SSH host key as a fingerprint: confirm it once, and from then on AnythingMCP refuses a different key before the password is sent, which protects you from a server that pretends to be yours.
- Scan. Click Scan folder. AnythingMCP lists the folder without indexing it yet and shows a tree with the number of files in every subfolder, the file types and the years. Folders that look like archives (
Alt,Old,Archive,Backup) are already unticked, so Claude never quotes an outdated document. A bar compares your selection with your plan while you tick folders, file types or an age limit; when the selection is larger than the plan, untick more or upgrade. - Describe. AnythingMCP drafts the connector's description from the scan, for example "Exported reports and document archive of the accounting team, PDF and Excel, 2019–2026". This text tells Claude what the folder contains and when to search it. If an AI model is configured on the server, the wizard can draft it with AI and says exactly what is sent (folder names, file names, the first 2,000 characters of up to 20 sample files, personal data masked); otherwise the draft is built from the folder structure. Edit it as you like.
- Activate. Choose the tool prefix, the sync interval and the MCP servers that get the connector, and click Activate.
The first sync starts right away and runs in a separate background service: Claude and your other connectors keep answering at full speed while it runs. The connector's Files tab shows its progress: files found, indexed and skipped, the last and next sync, and a Sync now button.
Step 2 — Copy the server URL
In MCP Servers, open the server you chose in the wizard and copy its URL.
Step 3 — Add the connector in Claude
Your server URL: in AnythingMCP, open MCP Servers → the server this connector is on, and copy its URL (
https://cloud.anythingmcp.com/mcp/…). Use it wherever this guide showsYOUR_SERVER_ID.
Recommended path: works on claude.ai web without editing any config file.
- In Claude, open Customize → Connectors.
- Click +, then Add custom connector.
- Enter a name, e.g.
Storage Box, and the remote MCP server URLhttps://cloud.anythingmcp.com/mcp/YOUR_SERVER_ID, then click Add. - Click Connect and sign in to AnythingMCP (OAuth) to allow access.
The same connector appears in the Claude desktop app. In Claude Code: claude mcp add --transport http anythingmcp https://cloud.anythingmcp.com/mcp/YOUR_SERVER_ID --header "X-API-Key: <your MCP API key>".
Step 4 — Ask
Start a new chat and ask about the content of your files. Claude searches the index first, opens the pages that matched and answers with the file name, the page and the file's date, so you can check the source.
What Claude can do with the files
| Tool | What it does |
|---|---|
<prefix>_guide | How to search this folder |
<prefix>_search | Full-text search over every indexed page: the files with the best page, a snippet and the date; filters for folder, file type and dates, newest first on request |
<prefix>_find | Find files by name (typo-tolerant) or pattern |
<prefix>_list | Show one folder: subfolders with counts, files with size and date |
<prefix>_read | Read a file as text, by pages ("1-3,7") or in windows of characters |
<prefix>_grep | Exact search for codes and numbers with two lines of context; spreadsheet rows come with their column names |
<prefix>_status | Sync state, counts, pause reason and a summary of the content |
<prefix>_view_page | Shows one page of a PDF, or an image file, as a picture: drawings, stamps, photos and tables the text does not capture |
<prefix>_get_file | The file itself as a temporary download link for members of the workspace (images are attached too) |
<prefix>_share_file | A temporary link (internal or public); switched off at install |
<prefix> is the tool prefix you choose in the wizard, for example pricelists. Every tool except _share_file only reads.
Supported files: text is extracted from over 100 file types: PDF (page by page); Word, Excel and PowerPoint including macro and template variants and .xlsb; OpenDocument; Outlook .msg, .eml and .mbox with the text of their attachments; .rtf, .epub, .xps; Apple Pages, Numbers and Keynote (through their PDF preview); .svg and the text of .dxf CAD drawings; CSV, Markdown, HTML, XML, JSON, YAML, calendars and contacts; and source code. Every other file is still found by name, and PDF pages and images can be shown to the AI as pictures. Scanned PDFs without a text layer are read with OCR on Enterprise.
Prompts to try in Claude
- "Search the archive for the 2024 annual report and summarise the revenue figures."
- "Which exported reports mention Nordtor as a supplier?"
- "Find every Excel row with the cost centre 4711."
- "Give me the newest monthly export as a download link."
How syncing works
AnythingMCP compares the size and modification time of every file with the last run and downloads only new and changed files. Files whose content is identical to the indexed version are recognised by their checksum (SHA-256) and not processed again. Deleted files leave the index. A file that cannot be read is retried on the next syncs and, after three failures (at once for password-protected files), stays findable by name, with the reason on the Files tab.
Syncing runs in its own background service, separate from the server that answers Claude: a large first sync does not slow down your questions, and search keeps working with the last index while a sync runs.
When the folder outgrows your plan, the sync pauses before downloading anything instead of indexing only part of it: the Files tab shows how many files the folder now holds against your plan, Claude keeps searching the files already indexed and is told that newer files may be missing, and you can untick folders or upgrade. A paused folder is checked again once a day and resumes by itself when it fits.
Security
- Read-only. The connector only lists and reads. It has no tool that writes, renames or deletes on the Storage Box.
- Your workspace only. The index is stored per workspace and every query is limited to it, by the application and again by the database (Row-Level Security).
- Credentials are stored encrypted and decrypted only in the background sync service. Claude and your colleagues never see them.
- Who sees what follows your MCP roles: a role without the file tools does not see them. On Enterprise you can also allow or deny folders per role.
- File content is data. The tools tell the model never to follow instructions found inside a file.
- Sub-account. Limited to one folder, it keeps the rest of the Storage Box out of reach.
- Host key pinned. The SSH host key of the Storage Box is checked on every connection.
- AnythingMCP Cloud connects only to servers on the public internet: a host in a private network is refused, also after a redirect. A server in your office network must be reachable from outside, or you run AnythingMCP self-hosted next to it.
Limits by plan
| Trial (Cloud) | Starter | Team | Enterprise | |
|---|---|---|---|---|
| File Search connectors | 1 | 1 | 3 | by contract |
| Indexed files, all connectors together | 500 | 500 | 2,000 | by contract |
| Largest file | 10 MB | 25 MB | 50 MB | 200 MB |
| Automatic sync | daily | daily | daily | down to every 15 minutes |
| Manual syncs per day | 1 | 1 | 3 | unlimited |
| File Share (download links) | – | ✓ | ✓ | ✓ |
| OCR, folder permissions per role, workspace policy | – | – | – | ✓ |
Self-hosted: Starter includes one File Search connector with up to 500 files, Team three connectors with 2,000 files, Enterprise has no limit and syncs down to every 5 minutes; the largest file is set by FILES_MAX_FILE_BYTES (100 MB by default). Self-hosted File Search needs a paid licence or the 30-day trial, which has the Team limits; the free Community edition has no File Search after the trial. All plans.
FAQ
Does it work on the Claude Free plan? Yes. Free plans can add one custom connector, and AnythingMCP counts as one no matter how many connectors you put on the server.
Why port 23? A Storage Box offers SFTP on port 23 once SSH support is enabled in its settings.
Main user or sub-account? A sub-account limited to one folder: it can be created in Robot under Storage Box → Sub-accounts and keeps AnythingMCP out of everything else on the box.
Does AnythingMCP copy all of the Storage Box? No. AnythingMCP keeps the extracted text, not the files. A file is read again only when it changed, or when you ask for the file itself.
Is vector search used? No. Company documents are searched for exact names, codes and numbers, where keyword search with page-level results is more reliable than embeddings.
Related
- Search your company file server with AI
- Find prices in supplier price lists with AI
- Search inside Nextcloud and WebDAV files with AI
- Share files from an AI chat as download links
- Connect a Hetzner Storage Box to ChatGPT
- Connect a Hetzner Storage Box to Gemini
- Connect a Hetzner Storage Box to Microsoft Copilot
- Connect a Hetzner Storage Box to Cursor
- Connect a Hetzner Storage Box to Muse
- Other file sources in Claude: SFTP server, FTP / FTPS server, WebDAV server, S3 bucket, Nextcloud, ownCloud, Synology NAS, QNAP NAS, Fritz!Box storage, Network folder (self-hosted)
Was this guide helpful?