AI Humanizer — Humanize AI Text

AI Text Humanizer — Write Naturally

Give stiff or repetitive drafts a more natural flow. Paste your text, choose how you want to rewrite it, and check the price before you start. You can work on one paragraph or a batch of drafts, then download the results for a final read-through.

Use it for emails, articles, product copy and other writing you are authorized to edit, including drafts from ChatGPT or Claude. The rewrite comes with a comparison of numbers, length and wording to help you decide what needs another look. Read the result in context before sharing it.

Find the right fit

Mode	What it is for	API value
Humanizer v4	Recommended for natural phrasing and varied sentences, with output length close to the original draft.	premium
Meaning Preserver	Prioritizes the original meaning and a similar length. Suited to technical writing, facts and figures.	precision
Humanizer v3	Rewrite in a creative, journalistic or professional voice. Choose the style that fits your draft.	balanced
Undetectable Pro	Rewrites AI-generated drafts with AI detection checks in mind.	deep
SEO Model	Rework blog posts and website copy with keyword-aware, conversational phrasing.	seo

Humanizer v3 offers four style settings: default, creative, journalistic and professional. The other modes use their default style. Leave language set to auto to keep the draft's language, or enter an output language such as English.

All five modes accept up to 12,000 characters per text. Longer drafts can be split into separate texts.

Pricing, before you begin

Pay per event, with platform usage included. The live Pricing tab is the source of truth for current prices. Each run has a $0.0025 start fee, plus a charge for each successfully saved, distinct rewrite.

Rewrite	Per result
Humanizer v4, Meaning Preserver, Humanizer v3 or Undetectable Pro: up to 20 words	$0.10
The same four modes: 21–100 words	$0.16
The same four modes: over 100 words	$0.70
SEO Model, at any length	$0.70

Words are counted by whitespace in the trimmed draft. Submit up to 20 texts per run. Repeated drafts, after trimming surrounding whitespace, share one result and one result charge. The current Pricing tab shows the effective prices; account subscription fees and taxes follow your Apify terms.

One non-SEO rewrite of up to 20 words costs $0.1025 including startup. A batch of 20 distinct drafts of up to 20 words each costs $2.0025. A 21–100-word non-SEO rewrite costs $0.1625 including startup.

To check the price first, select Preview price under Action (operation: "quote"). The preview shows distinct-text counts, per-item event names and prices, duplicate input positions and the complete rewrite total. A preview costs the $0.0025 start fee and sends no text to the rewriting service. A later rewrite is a separate run with its own start fee. Previewing and rewriting one non-SEO draft of up to 20 words therefore costs $0.105 in total.

Quick start: humanize your text

{
  "text": "Our team makes practical tools for busy people. We listen to customer feedback and test each change carefully.",
  "operation": "quote",
  "mode": "balanced",
  "style": "professional",
  "language": "auto",
  "additionalTexts": [],
  "includeOriginal": true
}

Check OUTPUT.quote.totalChargeUsd, then choose Humanize text under Action (operation: "humanize") to rewrite the same drafts. Add up to 19 more strings in Additional texts (additionalTexts). They use the same mode, style and language. Turn on Include original text (includeOriginal) to keep each original beside its rewrite for comparison.

Workflow: from draft to finished copy

For example, a content team can collect several introductions, preview the batch price in Humanizer v3, then rewrite with a professional style. The returned input positions keep each result connected to its draft, even when the batch contains duplicates. An editor can review flagged results before exporting the approved copy.

Output: what comes back

Each distinct rewrite is saved as a Dataset row. The original input positions stay attached, so the results are easy to match to your drafts.

Field	What it tells you
humanizedText	Your rewritten text
inputIndexes	Which submitted drafts share this result
originalText	Your original, when requested
mode, style, format	The selected settings; format is plain text
inputWordCount, outputWordCount	Length before and after rewriting
readabilityScore	A readability value when available, otherwise null
preservation	Length ratio and digit-based figures kept, lost or added
needsReview, reviewReasons	Checks for changed figures, large length changes or unchanged output
billing	The result event, USD price and input word count; startup is separate
id, processedAt, model	Result ID, processing time and model code

An example saved result:

{
  "billing": {
    "event": "tiny-text-humanized",
    "priceUsd": 0.1,
    "wordCount": 8
  },
  "needsReview": true,
  "reviewReasons": [
    "LENGTH_CHANGED"
  ],
  "id": "18b2eeb85404b746f1b6774c69964700faaa0c747eab51a5d79da158a0451adc",
  "inputIndexes": [
    0
  ],
  "model": "ft_clean-v3",
  "mode": "precision",
  "format": "text",
  "style": "default",
  "preservation": {
    "wordCountRatio": 1.375,
    "figuresKept": [],
    "figuresLost": [],
    "figuresAdded": [],
    "unchanged": false
  },
  "humanizedText": "The library offers workshops where neighbors can learn new skills together.",
  "inputWordCount": 8,
  "outputWordCount": 11,
  "readabilityScore": 57.23363636363638,
  "processedAt": "2026-10-15T08:05:05.641Z",
  "originalText": "The library helps neighbors learn useful skills together."
}

These checks are a starting point for review. Compare names, quotations, figures and meaning with your draft, even when a result has no flags. If you also need a separate classification score, the AI Content Detector offers that workflow with its own charges.

For API and AI agent workflows

The API mode values in the table remain stable. Display names help people choose; use premium, precision, balanced, deep and seo.

curl -X POST 'https://api.apify.com/v2/acts/khadinakbar~ai-text-humanizer/runs?maxTotalChargeUsd=0.01' \
  -H "Authorization: Bearer $APIFY_TOKEN" \
  -H 'Content-Type: application/json' \
  -d '{"text":"The library helps neighbors learn useful skills together.","operation":"quote","mode":"premium"}'

Poll the returned run ID until it finishes, then read OUTPUT using its defaultKeyValueStoreId. For the example rewrite, use operation: "humanize" with an allowance of at least $0.1025. Read the saved rows through defaultDatasetId with the Apify Dataset API.

Agent workflow: Preview the authorized drafts with operation quote. Inspect the total price and confirm the spending allowance. Rewrite the same inputs, read OUTPUT and RUN_SUMMARY, and return the saved Dataset rows with their input positions and billing details. Send flagged results for review and explain a partial outcome before submitting remaining work.

Hosted MCP setup and public discovery are assessed separately from the regular Actor API.

Best results and recovery

Start with coherent plain-text paragraphs. Extract the prose from formatted documents before submitting it. Keep unrelated drafts separate so each rewrite has enough context.

The Actor checks the spending allowance before each request and processes texts one at a time. Each provider request has a 90-second deadline and is never retried automatically. If a response is incomplete or a request fails, the start fee still applies; a successfully saved rewrite receives its result charge. Completed results stay available when a batch stops early.

OUTPUT and RUN_SUMMARY explain the outcome: COMPLETE, PARTIAL, INVALID_INPUT, UPSTREAM_FAILED or CONFIG_ERROR. A saved checkpoint protects completed work. An uncertain write or charge needs reconciliation before replay. For help, share the safe error code and run ID with support, keeping confidential prose in authorized storage.

Your text and privacy

Rewriting sends your draft to an external processing service through an encrypted, owner-managed credential. Price previews calculate locally. Apify retains submitted INPUT and generated output according to the account's storage settings. Results stay scoped to their run; there is no shared content cache or external telemetry.

Use text you are authorized to process, follow applicable laws and terms, and check attribution and disclosure requirements before publication.

Builder's note

My goal was to make a rewrite easy to review and easy to budget for. I added price previews, grouped duplicate drafts and put the result charge beside each rewrite so people and automated workflows can see what happened. The review flags help you choose what deserves another read.