Made By Humans
API docsSign inSign up

Learn

Tutorials & Guides

Step-by-step instructions for every feature. From your first text check to full API integration and compliance reporting.

Quick start

New to Made By Humans? Start here.

  1. 1Create a free account with your name, email, and publication.
  2. 2Log in to reach the Dashboard — your main tool interface.
  3. 3Paste text in the right panel, or upload an image in the left panel.
  4. 4Click “Analyse” and review your results. Each analysis has a fixed token cost shown before you confirm.

All tutorials

Click any tutorial to expand the full step-by-step guide.

Check text for AI generation

beginner

Paste article text and get a sentence-level AI detection report with overall probability score.

Cost: 2 tokens

  1. 1Log in to your account and navigate to the Dashboard.
  2. 2In the Text Analysis panel (right side), paste your article text into the text area. You can also use 'Extract from URL' to pull text from a published article automatically.
  3. 3Optionally click 'Clean text' to remove formatting artefacts before analysis (free, no tokens).
  4. 4Click 'Analyse text'. The system will run AI detection and return results within a few seconds.
  5. 5Review the overall AI probability score at the top. Scores above 50% warrant closer editorial review.
  6. 6Scroll through the highlighted text — sentences flagged as likely AI-generated will be colour-coded by severity.
  7. 7Use the results to inform your editorial decision. The score supports judgement — it doesn't replace it.

Check an image for AI generation

beginner

Upload a photo or paste an image URL to detect if it was AI-generated or manipulated.

Cost: 2 tokens (+ 5 for heatmap)

  1. 1Log in and go to the Dashboard.
  2. 2In the Image Analysis panel (left side), either drag-and-drop an image or paste a direct image URL.
  3. 3Click 'Analyse image'. The image will be uploaded to secure storage and submitted for classification.
  4. 4Wait for the result (usually 5–15 seconds). You'll see a verdict: 'Human' or 'Artificial' with a confidence percentage.
  5. 5If the image is flagged as potentially AI-generated, click 'Request heatmap' (5 additional tokens) to see exactly which regions show manipulation signals.
  6. 6The heatmap overlay highlights suspicious areas in red/warm colours. Cool/blue areas appear authentic.
  7. 7For further verification, run 'Reverse image search' (2 tokens) to check if the image appears elsewhere online.

Verify image originality with reverse search

beginner

Check if an image appears elsewhere on the web and trace its origin.

Cost: 2 tokens

  1. 1After uploading an image for analysis, click the 'Web Detection' button on your submission.
  2. 2The system queries Google's web detection service to find matching and similar images across the internet.
  3. 3Review the results: full matches (exact copies), partial matches (crops or edits), and visually similar images.
  4. 4Check 'Pages with matching images' to see which websites have used this image before.
  5. 5Look at 'Web entities' for automatic labelling — this tells you what Google thinks the image depicts.
  6. 6Use these results to verify: Is this a stock photo? Has it been used in misleading contexts before? Is it original?

Run a forensic editorial review

intermediate

Get an AI-powered pre-publication checklist identifying unverified claims, unsourced statistics, and logical issues.

Cost: 3 tokens

  1. 1Navigate to the Dashboard and select the Editorial Interrogator (available in the text analysis tools).
  2. 2Paste the full article text. Optionally include the article title for better context.
  3. 3Click 'Interrogate'. The system sends the text to a forensic AI reviewer (not a rewriter).
  4. 4Review the structured report: it includes a risk level (low/medium/high), unverified claims, unsourced statistics, logical issues, and boilerplate flags.
  5. 5Each flagged item includes the exact quoted text, the concern, and a suggested verification action.
  6. 6Use the 'Verification questions' section as a checklist — these are specific questions to ask the author before publication.
  7. 7The interrogator does NOT rewrite or improve the text. It identifies risks and gives you a checklist.

Identify which AI model wrote the text

intermediate

Analyse writing patterns to estimate the likely AI model and detect mixed human/AI authorship.

Cost: 3 tokens

  1. 1From the Dashboard, access the Authorship Fingerprinting tool.
  2. 2Paste at least 200 characters of text (longer text gives more accurate results).
  3. 3Click 'Analyse fingerprint'. The system runs both AI detection and style pattern analysis.
  4. 4Review the likely model estimation — it will suggest which AI (ChatGPT, Claude, Gemini, etc.) may have generated the text.
  5. 5Check 'Mixed authorship' — if different sections have different styles, the tool flags where the style shifts.
  6. 6Review style metrics: lexical diversity, sentence length variance, transition word density, and paragraph uniformity.
  7. 7Use the interpretation notes for a plain-language summary of what the patterns suggest.

Check a video for deepfakes

intermediate

Submit a video URL for AI deepfake detection.

Cost: 8 tokens

  1. 1Ensure your video is accessible via a public URL (MP4, WebM, or MOV format).
  2. 2Use the Video Detection endpoint or tool in the Dashboard.
  3. 3Paste the video URL and submit for analysis.
  4. 4Wait for processing (video analysis takes longer than images — typically 30–60 seconds).
  5. 5Review the verdict: 'Human' or 'Artificial' with a confidence score.
  6. 6For videos flagged as potential deepfakes, consider seeking the original source material for comparison.

Verify audio authenticity

intermediate

Check audio recordings for synthetic or cloned voices.

Cost: 4 tokens

  1. 1Ensure your audio file is accessible via a public URL (MP3, WAV, M4A, or OGG).
  2. 2Use the Audio Verification endpoint or tool in the Dashboard.
  3. 3Submit the audio URL for analysis.
  4. 4Review the result: verdict on authenticity plus clone detection indicator.
  5. 5If flagged as synthetic, investigate the source. Voice cloning is increasingly common in fraud and misinformation.

Analyse image metadata and provenance

intermediate

Extract EXIF data to verify camera info, detect editing software, and check for C2PA credentials.

Cost: 3 tokens

  1. 1After uploading an image, click 'Metadata forensics' on the submission.
  2. 2The system extracts all available EXIF metadata from the image file.
  3. 3Review camera information: make, model, lens. Real photographs typically have this data.
  4. 4Check for editing software: Photoshop, Lightroom, GIMP indicate the image has been processed.
  5. 5Look for AI generation software: DALL-E, Midjourney, Stable Diffusion in metadata is a clear indicator.
  6. 6Check the 'Suspicious indicators' section — stripped metadata, missing camera info on large files, or AI tool fingerprints.
  7. 7C2PA/Content Credentials status shows if the image has a cryptographic provenance chain (Adobe's standard).

Set up feed monitoring

advanced

Monitor RSS feeds for AI-generated content and get alerts when suspicious articles appear.

Cost: 1 token per article scanned

  1. 1Navigate to the Watchdog section or use the API endpoint.
  2. 2Add RSS feed URLs to monitor (up to 10 per account). Give each a descriptive label.
  3. 3Feeds can be your own publication, competitor sites, or freelancer portfolios.
  4. 4Trigger a scan manually or set up automated scanning via the API with a cron job.
  5. 5The system fetches the latest articles from each feed and runs AI detection on the text.
  6. 6Review results: articles with high AI probability are flagged for your attention.
  7. 7Use this to catch AI content in your own publication (post-publish audit) or monitor competitor practices.

Scan a publication archive

advanced

Submit a sitemap or URL list to bulk-scan your existing content for AI writing.

Cost: 1 token per article

  1. 1Prepare either a list of URLs or your sitemap URL (e.g., https://yoursite.com/sitemap.xml).
  2. 2Submit to the Bulk Scanner with either { urls: [...] } or { sitemapUrl: '...' }.
  3. 3The system processes up to 50 URLs per request, extracting article text from each.
  4. 4Each article is scanned for AI-generated content using the same detection models.
  5. 5Review the summary: total processed, failures, and a per-URL breakdown with AI scores.
  6. 6Articles with fractionAi above 0.5 are flagged. Sort results by score to prioritise review.
  7. 7For large archives, submit multiple batches. Token cost is just 1 per article at bulk rates.

Generate a compliance report

advanced

Export verification activity for audit, regulatory, or internal compliance purposes.

Cost: Free

  1. 1Navigate to Account → Compliance or use the API endpoint /api/compliance/report.
  2. 2Set the date range for your report (defaults to last 30 days).
  3. 3Choose format: JSON for programmatic use, CSV for spreadsheet analysis.
  4. 4The report includes: total checks performed, breakdown by type, detection results, and all verification badges issued.
  5. 5Each submission is listed with date, type, who checked it, result, and confidence score.
  6. 6Use this for EU AI Act disclosure compliance, internal audits, or demonstrating due diligence to stakeholders.
  7. 7Reports can be generated on-demand or scheduled via API for regular automated exports.

Integrate via REST API

advanced

Set up API authentication and make your first programmatic verification request.

Cost: Varies by operation

  1. 1Go to Account → API Keys and create a new API key. Give it a descriptive label.
  2. 2Copy the key immediately — it's shown only once.
  3. 3Get a bearer token: POST /api/v1/auth/token with your apiUserId and apiKey.
  4. 4Use the bearer token in subsequent requests: Authorization: Bearer <token>
  5. 5Submit text for analysis: POST /api/text/pangram with { text: '...' }
  6. 6Submit images: First upload to get a submission ID, then POST /api/vision/identifai/submit
  7. 7Poll for results: GET /api/vision/identifai/poll?submissionId=...
  8. 8See the full API documentation at /api-docs for all available endpoints and parameters.

Ready to start verifying?

Create an account and run your first check in under a minute.

Create free accountAPI documentation