Pseudonym MCP runs on your own machine — the client starts it, so there is no endpoint to ping. 106 installs a week from npm. Last commit 4 Sep 2026.
Locally pseudonymizes sensitive text before cloud LLM work and restores the tokens afterward.
Today is the operative word: we check Pseudonym MCP every 15 minutes and re-read its code on every release. Watch it and you find out the day that stops being true.
This server runs on your own machine — install it with the package manager and the client starts it for you. Package name taken from the official registry entry.
claude mcp add pseudonym-mcp -- npx -y pseudonym-mcp
{
"mcpServers": {
"pseudonym-mcp": {
"args": [
"-y",
"pseudonym-mcp"
],
"command": "npx"
}
}
}
[mcp_servers.pseudonym-mcp]
command = "npx"
args = ["-y", "pseudonym-mcp"]
{
"mcpServers": {
"pseudonym-mcp": {
"args": [
"-y",
"pseudonym-mcp"
],
"command": "npx"
}
}
}
{
"mcpServers": {
"pseudonym-mcp": {
"args": [
"-y",
"pseudonym-mcp"
],
"command": "npx"
}
}
}
Pseudonymizes PII before your LLM and returns a cryptographically signed receipt per response.
Pseudonymise and restore patient identifiers & PII in text — local, HIPAA Safe Harbor mode.
Clean raw HTML into LLM-ready text before agents spend tokens.
Web grounding for local & cloud LLMs: search, fetch, token-capped evidence packs, and research.
Context GC for LLM agents: offload large tool outputs and recall them to save tokens.
Agents report which MCP servers, APIs and CLIs actually worked. Look up before use, report after.
Price a Claude, GPT, Gemini or Grok call before you send it. Runs locally.
Redacts PII/PHI from text before it reaches an LLM. Fail-closed, self-hosted, no egress.
Answers built from our own checks of this server.