humanizer-ru · vladimir-human — Claude Code MCP server: what it does, how to install
humanizer-ru strips bureaucratic phrasing, templates and AI tells from Russian text. Runs offline, keeps numbers and names intact.
MIT ★ 153 updated 2026-10-21
claude mcp add humanizer-ru -- uvx humanizer-ruWhat humanizer-ru does
The server edits Russian text after a neural model: removes bureaucratic turns, template phrases and ChatGPT-style formatting. Numbers, dates, names and links stay unchanged — the bundled check_facts.py script verifies that. A second script, clean_paste.py, cleans chat paste artifacts: contentReference markers, utm_source=chatgpt.com tails and invisible characters.
In tests on 24 fresh texts from eleven models, humanizer-ru ranked first out of four participants (average rank 2.38 out of 5). Three models from other labs judged the results blind. Out of 24 texts, the server broke meaning in two cases; competitors broke it in two, nine and eight.
Runs via uvx; no need to install Python or dependencies. In Claude Code, add with claude mcp add humanizer-ru -- uvx humanizer-ru; in Cursor and Copilot, add through a JSON config with uvx and humanizer-ru in args.
What it helps with
- strip bureaucratic phrasing and templates from Russian text after a neural model
- preserve numbers, dates and names when an agent edits text
- check for lost numbers and links between source and edited version
- clean chat paste artifacts and invisible characters
Example prompts for the agent
- Make this Telegram post sound human
- Check what AI output looks like in this text, but don't edit
- Edit
README.md, leave code and links as is
How to install humanizer-ru
Platform CLI
claude mcp add humanizer-ru -- uvx humanizer-ruuvx runs the command named after the package; if the README names another one, use uvx --from humanizer-ru <command>
Platform CLI
codex mcp add vladimir-human-humanizer-ru -- uvx humanizer-ruuvx runs the command named after the package; if the README names another one, use uvx --from humanizer-ru <command>
Config file
{
"mcpServers": {
"vladimir-human-humanizer-ru": {
"command": "uvx",
"args": [
"humanizer-ru"
]
}
}
}uvx runs the command named after the package; if the README names another one, use uvx --from humanizer-ru <command>
Config file
{
"servers": {
"vladimir-human-humanizer-ru": {
"type": "stdio",
"command": "uvx",
"args": [
"humanizer-ru"
]
}
}
}uvx runs the command named after the package; if the README names another one, use uvx --from humanizer-ru <command>
Platform CLI
gemini mcp add vladimir-human-humanizer-ru uvx humanizer-ruuvx runs the command named after the package; if the README names another one, use uvx --from humanizer-ru <command>
Step by step, with commands for each platform: How to add an MCP server to Claude Code, Codex, Cursor, VS Code, Gemini CLI and more
Questions about humanizer-ru
How do I install humanizer-ru in Claude Code?
Run claude mcp add humanizer-ru -- uvx humanizer-ru in the terminal. The server starts via uvx; no need to install Python packages manually.
What does humanizer-ru do?
It edits Russian text after a neural model: removes bureaucratic phrasing, templates and ChatGPT-style formatting. Numbers, dates, names and links stay intact.
Which scripts are included in humanizer-ru?
check_facts.py compares source and edited text and shows missing numbers, dates, names and links. clean_paste.py strips chat paste artifacts — contentReference markers, utm_source tails and invisible characters.
Similar MCP servers
All MCP servers for Writing and docs →humanizer-ru by other authors comes first.