Writing-check API reference

Authentication

Every request carries your API key in the Authorization header. Keys are issued per account from Account → API, prefixed st_live_ (or st_test_), shown once at creation, and revocable at any time. Keep them server-side; never ship a key in client code.

Authorization: Bearer st_live_…

Credits & limits

API usage draws from the same credit balance as your Silvertext plan; a check costs the same credits whether it runs from the app, the API, or the MCP server. Requests are limited per key (burst and per-minute); throttled calls answer 429 with a Retry-After header. Submissions accept an Idempotency-Key header so a retried POST can never double-charge.

Submit a check

Send text or a file with the checks you want: plagiarism, grammar, citations, and AI content. The API answers immediately with a check id; most checks finish in about a minute, and a long document takes longer.

The document can arrive however your client sends it: JSON with text or a base64 file, a multipart upload, a form field, or the raw request body with any media type. Files are read by their bytes, never by their name or declared type: .txt, .md, .html, .csv, .docx, .xlsx and .pdf are read natively, scanned PDFs and JPEG/PNG/GIF/WebP images by OCR, and text is decoded from UTF-8, UTF-16 or Windows-1252. Size, type, emptiness, word limits and language are all judged before the check is accepted, so a 202 means the document was read and will run; there is nothing to convert on your side first.

Documents can be prose, source code, or both, in the language they were written in. The pipeline detects each segment's language and searches accordingly, and the report's languages and contentType fields say what it found. Languages written without spaces between words (Chinese, Japanese, Thai) aren't supported yet: those are refused at submission with code unsupported_language, and a refused or rejected check never consumes credits.

const res = await fetch('https://api.silvertext.com/v1/checks', {
  method: 'POST',
  headers: {
    Authorization: 'Bearer ' + process.env.SILVERTEXT_API_KEY,
    'Content-Type': 'application/json',
  },
  body: JSON.stringify({
    title: 'Q3 launch post',
    text: 'Your document text…',
    options: { plagiarism: true, grammarSpelling: true,
               citations: false, aiContent: true, style: true },
  }),
});
const { id } = await res.json();

Sending a file

import { readFile } from 'node:fs/promises';

// Inside JSON: base64 (what an MCP tool or an agent hands over)
const bytes = await readFile('thesis-draft.docx');
const res = await fetch('https://api.silvertext.com/v1/checks', {
  method: 'POST',
  headers: {
    Authorization: 'Bearer ' + process.env.SILVERTEXT_API_KEY,
    'Content-Type': 'application/json',
  },
  body: JSON.stringify({
    file: bytes.toString('base64'),
    filename: 'thesis-draft.docx',
    options: { plagiarism: true, aiContent: true },
  }),
});
const { id } = await res.json();

Response · 202

{ "id": "9b2f1c34-7c1d-4c7e-9a44-2f6f4be8d915" }

Get a check

Status while running, the report when done. You control the payload size: ?view=summary returns scalars and scores only, and ?include=scores,issues,text,matches,sources,spans,meta cherry-picks exactly the report sections you want, from a bare score to the full annotated document.

let check;
do {
  const res = await fetch(
    'https://api.silvertext.com/v1/checks/' + id + '?view=summary',
    { headers: { Authorization: 'Bearer ' + process.env.SILVERTEXT_API_KEY } },
  );
  check = await res.json();
  if (check.status !== 'done') await new Promise((r) => setTimeout(r, 3000));
} while (!['done', 'failed', 'rejected'].includes(check.status));

Response · 200 (view=summary)

{
  "id": "9b2f1c34-7c1d-4c7e-9a44-2f6f4be8d915",
  "status": "done",
  "title": "Q3 launch post",
  "wordCount": 1184,
  "scores": {
    "plagiarism": 14,
    "grammarSpelling": 2,
    "citation": 0,
    "aiContent": 9,
    "clarity": 6
  },
  "issueCounts": { "grammar": 3, "citations": 0, "clarity": 4 },
  "reportUrl": "https://app.silvertext.com/reports/9b2f1c34…"
}

List checks

Recent checks with status, scores, and open-issue counts. Useful for dashboards and for agents resuming work.

const res = await fetch('https://api.silvertext.com/v1/checks', {
  headers: { Authorization: 'Bearer ' + process.env.SILVERTEXT_API_KEY },
});
const { checks } = await res.json();

Cancel a check

Stops a queued or running check. Cancelled work releases its credit reservation; only completed checks settle usage.

await fetch('https://api.silvertext.com/v1/checks/' + id + '/cancel', {
  method: 'POST',
  headers: { Authorization: 'Bearer ' + process.env.SILVERTEXT_API_KEY },
});

Usage & balance

The credit balance and current period. Check it before large batches, or read the same numbers from any scan response.

const res = await fetch('https://api.silvertext.com/v1/me/usage', {
  headers: { Authorization: 'Bearer ' + process.env.SILVERTEXT_API_KEY },
});
const usage = await res.json();

Response · 200

{
  "creditsRemaining": 236,
  "creditsUsed": 4,
  "periodStart": "2026-08-04T00:00:00Z",
  "periodEnd": "2026-09-04T00:00:00Z"
}

Errors

Errors are JSON with a stable machine-readable code and a human-readable message. Branch on the code; show the message.

{
  "error": {
    "code": "insufficient_credits",
    "message": "This check needs 12 credits; 3 remain. Top up from your account."
  }
}