Anonymize Mcp
by Buggy1111communityv0.10.2
2
MCP server for Czech and multilingual NLP tasks from LINDAT/ÚFAL, including entity extraction, anonymization, morphology analysis, readability checking, spell correction, and translation. It can run anonymization in a local zero-egress mode using an on-device NameTag model.
From the registry: Anonymize PII & redact text (GDPR): Czech + 35-lang NER, morphology, translation, spellcheck
Install & configure
Prerequisites
- requiredInstall anonymize-mcp from PyPIInstall the Python package so the `anonymize-mcp` command is available.
https://pypi.org/project/anonymize-mcp/$ pip install anonymize-mcp - optionalOptional: install local/offline anonymization extrasFor zero-egress local anonymization mode, install the optional `local` extras to add the local NameTag backend.
https://pypi.org/project/anonymize-mcp/$ pip install "anonymize-mcp[local]" - optionalOptional: download local NameTag modelIf using zero-egress local anonymization mode, run the local backend once to download the model (~31 MB).
$ python -m anonymize_mcp.local_backend
Install in your client
Paste into the AI you're chatting with — it will figure out the rest
Please install the `anonymize-mcp` MCP server into my current AI client (that's you).
Required prerequisites (do these first if not already done):
- **Install anonymize-mcp from PyPI** — Install the Python package so the `anonymize-mcp` command is available. Run: `pip install anonymize-mcp` (https://pypi.org/project/anonymize-mcp/)
Optional prerequisites:
- Optional: install local/offline anonymization extras — For zero-egress local anonymization mode, install the optional `local` extras to add the local NameTag backend. Run: `pip install "anonymize-mcp[local]"` (https://pypi.org/project/anonymize-mcp/)
- Optional: download local NameTag model — If using zero-egress local anonymization mode, run the local backend once to download the model (~31 MB). Run: `python -m anonymize_mcp.local_backend`
Canonical MCP server config (stdio transport):
- command: `anonymize-mcp`
- args: []
- optional environment variables:
- `ANONYMIZE_MCP_LOCAL`: Set to `1` to enable zero-egress local anonymization mode for the `anonymize` tool. (example: `1`)
- `ANONYMIZE_MCP_NAMETAG_MODEL`: Path to a local NameTag model file to use in local anonymization mode instead of the auto-downloaded cached model. (example: `<path-to-model.ner>`)
- `ANONYMIZE_MCP_NO_DOWNLOAD`: Set to `1` to disable automatic model download in local anonymization mode; you must provide the model manually. (example: `1`)
Note: Default mode uses public external APIs at `lindat.mff.cuni.cz` and `quest.ms.mff.cuni.cz`. The server works without any environment variables in cloud mode. Zero-egress local mode only applies to the `anonymize` tool; other tools still call cloud APIs.
Add this MCP server to my current client's config in the correct format for you. If you need secrets or credentials I haven't provided, ASK me — do not invent values or leave raw placeholders. After adding it, tell me how to verify the server is connected.Note: Default mode uses public external APIs at `lindat.mff.cuni.cz` and `quest.ms.mff.cuni.cz`. The server works without any environment variables in cloud mode. Zero-egress local mode only applies to the `anonymize` tool; other tools still call cloud APIs.
Tools (6)
extract_entities
Multilingual named entity recognition with NameTag 3; Czech plus 34 other languages with auto-detection.
anonymize
Production-grade pseudonymization/anonymization using regex pre-pass and NameTag; supports placeholder mode and local zero-egress mode.
analyze_morphology
Tokenization, lemmatization, POS tagging, and dependency parsing with UDPipe; auto-detects 35 languages.
check_readability
Checks Czech readability using PONK.
correct_text
Czech spell checking and automatic diacritics handling with Korektor.
translate_text
Translates between 8 languages using Charles Translator.
Environment variables
ANONYMIZE_MCP_LOCALrequiredEnables zero-egress local anonymization mode.Prerequisites
- • Python3
- • python3
Packages
anonymize-mcp
pypiv0.10.2transport: stdio
Details
- Language
- Python
- License
- NOASSERTION
- Released
- Jun 10, 2026
- Last commit
- Jun 10, 2026
- Registry name
- io.github.Buggy1111/anonymize-mcp
Categories
Related Servers
Labelhead Artist Momentum
Trending hip-hop artist momentum scores across four cultural dimensions.
74.7k
PostHog MCP Server
Official PostHog MCP Server for product analytics, feature flags, experiments, and more.
37.3k
Openmetadata Mcp
Official OpenMetadata MCP: governed context and business semantics for AI assistants and agents.
14.6k