Developers

Give your agent a writing check

Silvertext exposes the same plagiarism, grammar, citation, and AI-content checks the app uses, as four MCP tools. Your agent can run them before a draft ships.

MCP Server setup REST API

The loop

# Typical agent loop
1. silvertext_get_usage()
   → credits_remaining: 236
2. silvertext_submit_check(text: "<draft>", wait_seconds: 45)
   → { status: "done",
       scores: { plagiarism: 3, grammarSpelling: 1,
                 citation: 0, aiContent: 12, clarity: 5 },
       report_url: "https://app.silvertext.com/reports/…" }
3. If the draft is clean, ship it.
   If not, apply the returned fixes and check again.

Same four tools in every MCP client. Credits come off one balance.

Three steps

Get an API Key

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

Request a key

Connect Your Client

Point Claude, ChatGPT, Gemini, Cursor, VS Code, or any MCP client at the Silvertext server. One URL and a Bearer header.

MCP Server Setup

Have the Agent Check the Draft

Call silvertext_submit_check. Most reports finish in that one call. Longer runs hand back a check id to poll with silvertext_get_check.

See the tools

What a good loop looks like

Read the balance first so you do not start a check you cannot finish. Submit the draft. Treat a clean report as a green light; treat findings as work to apply, then check again.

MCP or REST

Use the MCP Server when the caller is an agent that already speaks MCP. Use the REST API when you are wiring checks into your own product or a pipeline. Both hit the same engine and the same credits.