MCP server.
The same deterministic linter and humanizer that power Typographer, exposed as Model Context Protocol tools. Any MCP host can grade its own writing, see the machine tells, and rewrite until the prose reads human.
Install
Add the server to your claude_desktop_config.json, then restart:
{
"mcpServers": {
"typographer": {
"command": "npx",
"args": ["-y", "@701am/typographer-mcp"],
"env": { "TYPOGRAPHER_API_KEY": "tg_live_xxx" }
}
}
}It works the same in any MCP host that speaks the protocol. Point it at a different deployment with TYPOGRAPHER_API_URL (default https://701am.com).
Tools
| Tool | What it returns | Cost |
|---|---|---|
score_text(text, mode?) | The three pillars (voice, readability, information gain) and every tell found. | Free, deterministic |
find_ai_tells(text, mode?) | Just the machine tells, the fastest signal to act on. | Free, deterministic |
humanize(text, mode?) | A rewrite that strips the tells, with an adversarial self-revision loop. | Metered, calls a model |
Every tool takes an optional mode, one of these five: general, marketing, technical, academic, news.
The loop it enables
draft -> score_text -> find_ai_tells -> rewrite -> score_textBecause the score is a deterministic check and not another model's opinion, the agent can trust it instead of gaming it. That is the point: a verification primitive an agent can call, not a black box that grades on vibes.
API keys
Sign in at 701am.com and create an API key in your account, then set it as TYPOGRAPHER_API_KEY. The two read tools are free and deterministic. humanize counts against your plan.