Developers · MCP Server

Give your AI agent a complete writing check

Works with Claude, ChatGPT, Gemini, Cursor, VS Code, and any MCP-compatible agent. One URL, one API key, and your agent runs the same plagiarism, grammar, citation, and AI-content checks as the Silvertext app, before a draft ships.

Get started Agentic Guide

~/.cursor/mcp.json

{
  "mcpServers": {
    "silvertext": {
      "url": "https://api.silvertext.com/mcp",
      "headers": { "Authorization": "Bearer st_live_…" }
    }
  }
}

Give Your Agent the Full Check

The MCP server plugs Silvertext into Claude, ChatGPT, Gemini, Cursor, and any MCP-compatible agent as callable tools: submit a draft, get plagiarism, grammar, citation, and AI-content scores, and read your credit balance.

Agentic Guide

Build Checks into Your Product

Submit text or files and get back a structured report: plagiarism matches with sources, grammar fixes, citation issues, and an AI-content score. Choose a bare score or the full annotated document per request.

REST API

Gate Publishing Pipelines

Batch-check a content library or wire an originality gate into your CMS, Zapier or n8n flow. It uses the same credits as your plan, with no separate API pricing to learn.

What your agent can call

Submit, fetch, list, and a balance read. Every finished check carries scores for plagiarism, grammar, citations, and AI content, plus the credits it used and what remains, so agents can budget without guessing.

silvertext_submit_checkCredits per words checked

Submit text, a URL, or a file. One call runs plagiarism, grammar, citation, and AI-content checks. Long-polls up to ~55s so most reports come back finished; longer runs hand back a check id to poll.

silvertext_get_checkFree

Fetch a check by id: status while running, then the report itself. A one-line verdict, scores, top sources, and the full annotated document on request.

silvertext_list_checksFree

List recent checks with scores and open-issue counts, so an agent can pick up where it left off.

silvertext_get_usageFree

Read the credit balance and period usage before spending. Every scan response also carries credits_used and credits_remaining.

A check, from your agent's side

# What your agent's calls look like (same in every client)
silvertext_submit_check(text: "<draft>", wait_seconds: 45)
→ { status: "done", verdict: "No matching sources found.",
    scores: { plagiarism: 3, grammarSpelling: 1, citation: 0, aiContent: 12, clarity: 5 },
    report_url: "https://app.silvertext.com/reports/…",
    credits_used: 4, credits_remaining: 236 }

Get your API key

The same key works for the REST API and the MCP server, and usage draws from one credit balance.

Building against REST instead? Start on the API page. Prefer a walkthrough? Read the Agentic Guide.

What your agent gets

Works in Claude, ChatGPT, Cursor, VS Code, Gemini, and Codex
Four tools: submit, fetch, list, and usage
The same report and scores the Silvertext app produces
One key and one credit balance for REST and MCP