AI Humanizer API to Embed in Your SaaS Product
Ship a humanization feature your users never leave your app to use. A white-label REST API with per-word billing you can mark up into your own plans, an effective rate starting around $0.17 per 1,000 words, 40+ auto-detected languages, and an average response of about 20 seconds.
Word-based pricing from $0.17 per 1K words. White-label-friendly. 40+ languages.
Rewritten to keep your meaning and improve the human score, back in your UI in about 20 seconds.
Output
Built for SaaS products
A REST API you build into your own product and your own pricing, with predictable per-word cost.
A white-label call your users never see
Your backend holds the wh_live_ key and makes one POST to /v1/humanize, so the feature renders entirely inside your own UI under your own brand. No redirect, no branded result page, no leaking users to another domain.
Per-word cost you build into your own pricing
Billing is on input words against a fixed monthly allowance, not shifting credits, with an effective rate starting around $0.17 per 1,000 words. That maps cleanly onto a per-seat or per-word price in your product, so your margin holds as you scale.
Low latency that fits your UX
Average humanize response is about 20 seconds, with 40 requests per minute on Standard and 120 on Premium. Call it as a background job with a progress state and the rewrite never blocks the rest of your product.
A score you can gate on, not a black box
Premium returns 3 scored variations per request, each with a human-likeness score, so your app can auto-select the strongest, retry on low confidence, or surface quality. The same key also reaches POST /v1/detect for AI-authorship scoring inside your app.
One server-side call inside your own product
Your key never touches the browser. The feature lives in your UI, under your brand, with WriteHuman invisible to your users.
Your app
User submits a draft in your UI; your backend attaches the wh_live_ key server-side.
WriteHuman API
POST /v1/humanize rewrites the text and returns scored variations in about 20 seconds.
Your users
You return the top result under your own brand. WriteHuman never reaches the browser.
Add it to your backend in a few lines
Generate a wh_live_ key in the dashboard and store it as a server-side secret, holding up to 5 revocable keys so you can rotate credentials without redeploying your whole stack. Your app server forwards user text to POST /v1/humanize with an optional tone, then returns the humanized result to your front end under your own domain so WriteHuman stays invisible to end users. Read GET /v1/usage, which is never billed, to track words consumed, then map that draw onto the per-seat or per-word price you charge inside your own product.
// app/api/humanize/route.ts in your product.
// Your front end calls this; the wh_live_ key never reaches the browser.
export async function POST(req) {
const { text, tone = "professional" } = await req.json();
const res = await fetch("https://api.writehuman.ai/v1/humanize", {
method: "POST",
headers: {
Authorization: `Bearer ${process.env.WRITEHUMAN_API_KEY}`,
"Content-Type": "application/json",
},
// text must be 30 to 100000 chars. Billed on input words.
body: JSON.stringify({ text, tone }),
});
const data = await res.json();
// results is sorted best first. Return only what your UI needs.
return Response.json({ humanized: data.results[0] });
}Build it yourself, or ship it this week
Build humanization in-house
- Train, host, and tune your own rewrite model, then keep it current as detectors change
- Build language coverage, tone control, and an authorship scorer from scratch
- Carry unpredictable GPU and infra cost that is hard to price per user
- Own latency, scaling, and on-call for a feature that is not your core product
Call the WriteHuman API
- One POST /v1/humanize call, live the same day, no model to maintain
- 40+ auto-detected languages, 7 tones, and POST /v1/detect for scoring, in one API
- Per-word billing starting around $0.17 per 1,000 words you mark up into your plans
- Average 20-second responses with 120 req/min on Premium, scaling handled for you
Assistant-native product? Wire it up over MCP
If your product is built on AI assistants or agents rather than plain REST, connect WriteHuman over MCP instead of hand-rolling HTTP plumbing and shared keys. The connector uses OAuth 2.1 with PKCE, so there is no static secret to embed, and it works with Claude, Cursor, Codex, and any MCP client. It exposes humanize_text, detect_ai_text, and get_account, and it is included with Pro, Ultra, and API plans. See it at /mcp.
Explore the MCP connector- Three tools your agent calls directly: humanize_text, detect_ai_text, and get_account
- OAuth 2.1 with PKCE instead of shared keys, so there is no secret to rotate in client tooling
- Works with Claude, Cursor, Codex, and any MCP client, alongside the same REST API
Simple, word-based pricing
Pay for the words you send. Allowances reset every billing cycle, and top-up packs never expire.
Standard
$29 / mo- 125,000 words / month
- 40 requests / min
- Single humanized result per request
Premium
$69 / mo- 400,000 words / month
- 120 requests / min
- 3 variations with human-likeness scores
Frequently asked questions
How do I embed an AI humanizer in my SaaS so users never leave my app?
Call POST /v1/humanize from your own backend with your wh_live_ key, then return the humanized text to your front end under your own domain. It is a server-side REST API, so the feature lives entirely inside your UI and WriteHuman stays invisible to your end users. There is no redirect and no branded result page.
How is the API priced, and how do I build a predictable per-word cost into my own pricing?
Billing is on input words against a fixed monthly allowance: Standard is $29 for 125,000 words and Premium is $69 for 400,000 words, an effective rate starting around $0.17 per 1,000 words. Because the rate is per word rather than per credit, you can map it directly onto a per-seat or per-word price in your product. Non-expiring 125,000-word top-up packs cover overages without changing your base cost.
Can I white-label the humanizer so it appears as part of my own product?
Yes. Your backend holds the key and makes the call, so your users interact only with your interface and your brand. WriteHuman is never surfaced to them. Read GET /v1/usage, which is never billed, to meter consumption inside your own product.
How fast is the humanize endpoint, and should I call it synchronously?
Average humanize response is about 20 seconds, so we recommend an asynchronous call with a progress indicator rather than a blocking inline rewrite. Standard allows 40 requests per minute and Premium allows 120, with input up to 2,000 words per request, which lets you size a queue against your real traffic.
Is there a free way to try the API before I integrate it?
You can evaluate the endpoints and read the full request and response spec at /api/docs before committing, and pricing for the Standard and Premium plans is at /api#pricing. Start on Standard at $29 for 125,000 words a month to prototype your integration, then move to Premium for scored variations and higher rate limits once the feature ships.
How do I authenticate, and how many keys can I create and revoke?
Authenticate with a Bearer token using a key prefixed wh_live_ in the Authorization header. You can hold up to 5 keys per account and revoke any of them at any time from the dashboard, so you rotate credentials in production without forcing a full redeploy. Full request and response details are in the REST reference at /api/docs.
Build humanization into your product, not around it
One POST request, a per-word cost you can model, and output that reads as human, all inside your own UI. Read the full spec at /api/docs and word-based pricing at /api#pricing, then create a wh_live_ key and ship the feature.