Add Traceix to Claude, Cursor, Cline, Roo Code, or any MCP client. Then ask it to watch a folder, hash new files, search Traceix, and ask before uploading anything.
A file-check server your MCP client can use.
Copy the config, paste it into MCP settings, then paste the folder prompt.
Hash and review files before opening or uploading them.
Paste it into your MCP client once. After that, use the folder prompt whenever you want a check.
Check installers, archives, and random files before opening.
Watch new files from teams, clients, classes, or friends.
Hash changed files before running scripts or binaries.
Review malware or unknown files without uploading first.
File triage usually means hashing a file, checking reputation, scanning it when needed, and writing down the result. Traceix puts that workflow behind one MCP server.
Traceix handles lookup, scan status, and reporting. You keep control over uploads and final action.
Search SHA-256 first. If Traceix already has the file, use that result before uploading anything.
Use the same endpoint in Claude Desktop, Cursor, Cline, Roo Code, or your own MCP setup where supported.
Hash lookup, upload, classification, full scan, scan status, login, and API-key tools are exposed through MCP.
The result can live in Traceix instead of disappearing inside one chat thread.
Use this in any MCP client that supports a remote streamable HTTP server. Connect, list the tools, sign in if required, then run a hash lookup or file scan.
{
"mcpServers": {
"traceix-mesh": {
"type": "streamable-http",
"url": "https://mcp.traceix.com/mcp"
}
}
}These are the Traceix tools exposed through MCP. Some actions may require login, an API key, or available credits.
| Tool | What it lets agents do |
|---|---|
| traceix_register_user | Create a Traceix account. |
| traceix_user_login | Log in and receive a token. |
| traceix_user_get_api_key | Retrieve an API key for authenticated analysis. |
| traceix_search_sha_hash | Search Traceix by SHA-256. |
| traceix_upload_and_classify | Upload a sample for initial classification. |
| traceix_full_file_upload | Run deeper analysis and store results. |
| traceix_full_file_upload_status | Check scan jobs until complete. |
A Traceix Hub is planned for people who do not want every file check to depend on the cloud. It is meant to keep scanning closer to your own machine or network.
You do not need a Hub to use the MCP endpoint today. Hubs are for private local scanning and always-on local access when they ship.
Status: preorder / coming soon. Hubs are intended to be part of the Traceix Mesh network, but they are not represented here as finished or currently shipping.
Use a Hub if privacy is the main concern, or if you want local analysis available on your own network.
Hubs are for teams, classrooms, home labs, and sensitive workflows where private scanning matters more than cloud convenience.
The Hub is meant to sit nearby, handle local checks, and only escalate to cloud workflows when you choose.
Use native MCP when your client supports remote servers. If it does not, use an MCP bridge, an IDE client, or the Traceix API. Hubs are optional.
https://mcp.traceix.com/mcpCreate a remote MCP server and paste the Traceix endpoint.
Ask your client to list Traceix tools. Sign in when a tool requires it.
Search a hash, upload an approved file, check scan status, or write a summary.
Use this anywhere remote streamable HTTP MCP is supported.
{
"mcpServers": {
"traceix-mesh": {
"type": "streamable-http",
"url": "https://mcp.traceix.com/mcp"
}
}
}Add a custom MCP connection if your workspace exposes it. Use the server name and URL above.
Use the universal config in Claude's MCP setup, then restart Claude Desktop and confirm the Traceix tools appear.
Open MCP settings, add a remote server, paste the config, then search or scan files through Traceix.
Custom MCP support varies. If your app does not expose it, use an MCP-capable desktop client, automation layer, or wrapper.
Point your MCP client at the endpoint, request the tool list, sign in, then call the hash, upload, or status tool.
List the Traceix tools, then show me the workflow for searching a SHA-256 hash and uploading a file.
Start small. Check a hash or one folder first. Upload only after you approve the file.
{
"filename": "sample.bin",
"file_base64": "<base64 file content>",
"api_key": "<traceix api key>"
}Copy one of these into your MCP-connected client. Start with folder watch for everyday use.
For watching a folder and checking new or changed files.
Watch this folder: [PASTE FOLDER PATH]. When a file is added or changed, calculate its SHA-256 hash and search Traceix. Do not upload anything unless I say yes. Return a table with file name, path, SHA-256, Traceix result, and what I should do next.For checking known hashes before uploading anything.
Check these SHA-256 hashes in Traceix. First confirm each value is 64 hexadecimal characters. Return a table with hash, Traceix record found or not found, classification if available, scan status if available, and next step.For approved unknown files only.
Upload only the files I approve. Use the file name and base64 file content required by the Traceix tool, not a local path. After upload, check scan status until complete. Return classification, scan summary, indicators, and anything that needs manual review.For saving or sharing the result internally.
Write a short file-check report from the Traceix results. Include files checked, SHA-256 hashes, classifications, indicators, scan status, what is confirmed, what is uncertain, and the next step for each file.Add the MCP endpoint.
Pick folder watch, hash lookup, upload, or report.
Hash and search before uploading.
Upload approved files only.
Save findings and next steps.
A SHA-256 hash is 64 hex characters. Check the format before searching.
Do not put API keys in shared prompts or public repos.
Some scans take time. If you get a job ID, check status until it finishes.
Use the output as support, not as final proof.
Hash files, look them up, upload only what you approve, and save the result.