B2B Content Evaluation API

Your users upload things.
We tell you what they are.

Private, encrypted, server-side content evaluation. We analyze. We return JSON. We forget the image.

See Pricing View API Docs
🔒 Images never leave our encrypted pipeline
🗑️ Evaluate and forget — we don't retain content
JSON verdict in seconds
🛡️ Zero-compromise standards
🏛️ Private server LLM — not a cloud black box

Drop it in front of your upload pipeline.
That's the whole integration.

You send content. We evaluate it against clear, published standards. You get a JSON verdict. Your users never see the machinery.

01

You send us the content

POST a file, a URL, or base64 to our API. Text, image, or both. We accept anything your users can upload.

02

Ramon reviews it

Our private, server-hosted LLM evaluates against published standards. No black box. No mystery. You can read exactly what we check for.

03

You receive a verdict JSON

Flags, confidence scores, a soft message for your users, and a clinical description for your records. Structured. Consistent. Actionable.

04

We forget the content

Standard evaluation: no retention. We return the analysis and discard the source. You keep your data. We keep our promise.

Clear standards. Published openly.
No surprises.

Our evaluation criteria are not a secret. Here is exactly what Ramon checks for, and why.

🚫 Minor Detection

Any depiction of a person appearing under 18 — photographic, illustrated, animated, or AI-generated. Highest priority. Non-negotiable verdict.

🚫 Explicit Content

Sexual content of any kind. Returned as flagged with confidence score. Your app decides the threshold. We give you the signal.

🚫 Violence & Gore

Zero tolerance on our platform. Flagged immediately. Clinical description returned. Content not forwarded.

⚠️ Distress Signals

Crisis imagery, visible self-harm indicators, distress signaling. Flagged with optional forwarding to welfare monitoring. Your call.

⚠️ Hate Symbols

Recognized hate iconography, supremacist imagery, flags and insignia associated with hate movements.

ℹ️ Personal Information

Visible faces of real people, license plates, addresses, documents. Flagged for your privacy compliance review.

ℹ️ Age Verification (Batch)

Sort a directory by apparent age. Get a per-file report with confidence scores. Separate the content that needs human review.

ℹ️ Text & Tone

Mood analysis, spam detection, fraud signals, prohibited content in user messages. Same API, same JSON envelope.

We built this because
we needed it ourselves.

AIMentorAgency.com runs AI mentors who talk to real people about real things. We couldn't let harmful content into those conversations. So we built an evaluation layer.

Then we realized: everyone with a platform has this problem. Most of them are handling it badly — underpaid human moderators, or blunt automated tools that over-censor and infuriate users.

We're making our solution available. It runs on our hardware. Our standards are published. Our promise is simple.

  • We evaluate your content against published, transparent standards
  • Our LLM runs on private FileSpeeder hardware — not a third-party cloud
  • Standard evaluation: content is discarded after analysis
  • Flagged content held only when legally required — stated explicitly in our terms
  • You can read our evaluation criteria. All of them. Right now.
  • No training on your data. Ever.
  • No override switch for sale. No "enterprise jailbreak" tier.
  • No contract buys you different standards than everyone else gets.
  • No mystery. No black box. No "trust us."
🔬
Ramon
Chief Evaluation Officer
"I do not talk to your users. I talk to your platform. I read what they send. I return my verdict. I do not retain what I have seen — unless the law requires it, and then I keep it for the law, not for commerce."

Ramon is our private LLM evaluation engine. She has clinical precision and ethical standards that are not for sale. She runs on our hardware. Her criteria are public. Her findings are yours.

She is supervised by Sheila, who sets the standards she enforces. No B2B agreement, no API key, no payment tier changes what she considers acceptable. She has one set of standards and she applies them to everyone.

This is what you get back.

Structured. Consistent. Two descriptions — one for your records, one for your users. Every time.

// POST /api/scan/image → 200 OK
{
  "scan_id": "01HWXK9...",
  "verdict": "cleared", // cleared | flagged | blocked
  "negotiable": true,
  "flags": {
    "underage": false,
    "explicit": false,
    "violence": false,
    "distress": false,
    "hate_symbol": false,
    "personal_info": false
  },
  "confidence": 0.97,
  "descriptions": {
    "clinical": "Image depicts an adult female subject, approximately 30-40 years. No prohibited content detected.",
    "soft": "This image passed our standard content review.",
    "user_message": null
  },
  "retained": false,
  "scanned_at": "2026-05-28T14:32:00Z"
}

Simple. By the call.
No contracts required.

Buy tokens, use them, buy more. Or talk to us about a dedicated private instance — your data, your server, your peace of mind.

Starter

$0.01 / image scan

Token bundles starting at $10. No monthly fee. No commitment.

  • Image evaluation
  • Text evaluation
  • Full JSON response
  • No retention (standard)
  • isAuthorized.com login
Get API Key

Private Hosted

Custom

Your dedicated EvaluateContent instance. Your data never touches shared infrastructure.

  • Everything in Volume
  • Dedicated server instance
  • Your encryption keys
  • Air-gapped option available
  • Custom criteria review
Contact Us