The terms and conditions for using Tuteliq's child safety API, SDKs and dashboard, covering accounts, acceptable use, billing and liability. Read them in full.

Last updated: August 6, 2026 · Effective: August 6, 2026 · Version 2.1 (zero-storage model)

1. Introduction and Agreement

These Terms of Service ("Terms") govern your use of the services, website and API provided by Tuteliq AB (the "Company", "we", "us", "our"). By accessing or using our services, you agree to be bound by these Terms.

  • Service provider: Tuteliq AB
  • Swedish organization number: 5594811217
  • UK establishment: BR028608, 3rd Floor, Elite House, 60 Rodney Street, Liverpool, L1 9AD, United Kingdom
  • Website: tuteliq.ai
  • Support: support@tuteliq.ai

2. Services Description

2.1 What we provide

Tuteliq provides an AI-powered content detection API designed to identify and flag harmful content in real time, including:

  • Grooming and exploitation
  • Bullying and harassment
  • Scams and fraud
  • Self-harm content
  • CSAM (child sexual abuse material)
  • Other harmful content categories

2.2 Core service model: real-time processing, zero storage

What we do

  • You submit content (text, image, audio, video) via the API.
  • We process it in real time using AI detection.
  • We return encrypted incident reports with the detection results.
  • We immediately delete the original content.

What we don't do

  • We do NOT store your original content
  • We do NOT back up customer content
  • We do NOT retain data for training or analysis
  • We do NOT share customer data with third parties
  • We CANNOT decrypt the incident content we return, you hold the private key

2.3 What you receive

  • Detection result: flagged or safe determination
  • Confidence score: probability of accuracy (0–100%)
  • Incident details: type, severity, patterns detected
  • Mirroring payload: your own identifiers (timestamp, message ID, user ID) so you can match the incident to your content
  • Encryption: encrypted with the key pair generated in your browser, so only you can decrypt

Critical point: we return incidents; we do not return or store your original content.

2.4 Supported content types

  • Text messages and written content
  • Voice and audio data
  • Images and visual content
  • Video content
  • Documents

2.5 Service availability

  • Uptime target: 99.5% monthly uptime
  • Response time: 1.1-1.5s typical detection response
  • Maintenance: 24 hours' notice when possible
  • Support: email support 24/5

3. Eligibility and Account Requirements

3.1 Eligibility

  • Be at least 18 years of age
  • Have legal capacity to enter into a binding agreement
  • Comply with all applicable laws
  • Not be prohibited from using our services

3.2 Account creation

  • Accurate and complete information
  • A valid business email address
  • Organizational registration details
  • Payment information for paid plans

3.3 Account responsibility

  • Maintaining confidentiality of your private encryption key
  • All activity that occurs under your account
  • Notifying us of unauthorised access
  • Compliance with these Terms and applicable law
  • Lawful use of our services

4. Data Processing and Your Responsibilities

4.1 We don't store your data

Step

What happens

Data stored?

1. Submit content

You send content via encrypted API

No

2. Process

We run AI detection in real time

3. Generate incident

We create an encrypted incident report

4. Encrypt

Incident encrypted with your public key

5. Return

Incident sent back to you

6. Delete

Original content immediately deleted

You control everything

  • Your private key, Tuteliq never has access
  • Your incident decryption, only you can decrypt
  • Your incident storage, you decide how long to keep it
  • Your incident action, you decide what to do

4.2 Your legal responsibilities

Before submitting data

  • Have lawful authority to process the data
  • Determine your legal basis (contract, consent, legitimate interest)
  • Inform data subjects in your own privacy policy
  • Obtain necessary consents, especially for children

For children under 13 (COPPA)

  • Obtain verifiable parental consent before submitting any child data
  • Inform parents about Tuteliq's role
  • Provide parental access and deletion rights
  • Never market to children

For children under 16 (GDPR)

  • Obtain parental consent before processing
  • Provide transparency about processing

4.3 Data processing relationship

  • You are the data controller: you determine purposes and means.
  • We are the data processor: we process only as instructed by you.
  • Processing agreement: our Data Processing Agreement is mandatory for business customers.
  • Subprocessors: listed at tuteliq.ai/legal/subprocessors.

4.4 What you are responsible for

  • Deciding what to do with incident reports (delete, flag, warn, report to authorities)
  • Storing and managing your incident reports
  • Ensuring you have consent to process submitted data
  • Maintaining your private encryption key securely
  • Complying with the law when taking action on flagged content

4.5 Child safety and reporting obligations

Tuteliq is a detection tool, not a reporting platform. When our service flags potential harmful content (including CSAM or child exploitation), you are responsible for taking action and for reporting to the competent authorities in your jurisdiction. Many jurisdictions require prompt action on detected CSAM, typically within 24–48 hours.

  • USA: National Center for Missing & Exploited Children (NCMEC) CyberTipline, cybertipline.org
  • EU / UK: your national authority and the Internet Watch Foundation, iwf.org.uk/report
  • Other jurisdictions: your local law enforcement or designated authority

What Tuteliq does not do

  • Store any content you submit
  • Automatically report to authorities (this is your responsibility as the platform operator)
  • Retain detection history for our own purposes
  • Make reporting decisions on your behalf
  • Decrypt your incident reports (you hold the encryption key)

Tuteliq is a member of the Internet Watch Foundation (IWF) as part of our child safety commitment. Our zero-storage model ensures we never retain harmful content, and we provide detection capabilities to help you identify and act on it. For reporting guidance, see IWF (international), NCMEC CyberTipline (USA), or your national law enforcement agency.

Incident logging – customer control

You may enable or disable incident metadata logging at any time in your account settings, or override it per request via the API. Incident logging is enabled by default for new accounts and may be changed without penalty.

  • When enabled: Tuteliq stores encrypted incident metadata (detection results, confidence scores, detected tactics, mirroring payload) for your configured retention period (default 90 days, configurable from 7 days to 2 years), accessible via dashboard, API query and export.
  • When disabled: incident metadata is delivered in the API response only and deleted from Tuteliq systems immediately afterwards. You are solely responsible for retaining incident data if you need it; historical queries and exports are not available.
  • Content deletion applies regardless: the toggle affects metadata retention only. Original content is always deleted immediately after processing and is never stored or retrievable by you or Tuteliq.
  • Changes are not retroactive: a change takes effect within one minute for new API calls; incidents already stored remain accessible until they expire under your retention policy or you delete them.
  • Data subject rights: where logging is enabled, Tuteliq acts as processor and you as controller for incident metadata; we cooperate with verified requests within 30 days. Where logging is disabled, no incident data exists to disclose or delete.

4.6 Acceptable use, prohibited uses of Tuteliq services

Tuteliq's detection models, per-tactic breakdowns and confidence signals are proprietary. You may use our output to protect your own platform, but not to reconstruct, distil or compete with the service that produces it.

Prohibited: model replication and data distillation

  • Use incident reports, detection results, confidence scores or detection patterns returned by Tuteliq to create or train machine learning models
  • Collect Tuteliq output with the intent to distil, extract or reverse-engineer our detection capabilities, or to build datasets that replicate or approximate our models
  • Use Tuteliq to generate data for building a competing content detection system, harmful-content classifier or child-safety detection model
  • Reverse-engineer, decompile or otherwise attempt to derive the underlying logic, rules or model weights of our detection algorithms
  • Perform knowledge distillation from our API responses, or use per-tactic breakdowns to label or augment your own training datasets
  • Sell, license, share or publish incident data or detection patterns derived from our API for commercial purposes, except under a separate written research agreement

Permitted: legitimate platform operations

  • Incident analysis: reviewing individual incidents to decide moderation actions (remove content, suspend users, escalate to authorities).
  • Safety audits: analysing incident patterns within your own data for compliance reporting or safety reviews.
  • Threshold tuning: adjusting detection thresholds to your platform's false-positive and false-negative tolerance.
  • Custom workflows: building incident-routing logic specific to your platform.
  • Research (with approval): academic research on harmful-content detection under a separate research agreement.

These uses are legitimate because you are using Tuteliq as a detection tool to support your own judgement, not extracting its capabilities.

Consequences of violation

Breach of this section is a material violation of these Terms. Where we detect or reasonably believe you are collecting output to train competing models, performing distillation, reverse-engineering our detection logic or attempting to replicate our model, we may immediately suspend your account and access without refund for the remainder of the billing period, terminate this Agreement on written notice, and pursue legal remedies including injunctive relief.

Audit rights

  • Monitoring API usage patterns for signs of systematic data collection (for example high-volume requests with identical or synthetic content)
  • Auditing your systems on 10 business days' notice to verify compliance
  • Requesting written certification that you are not using Tuteliq output for model training or distillation
  • Requiring deletion of any dataset derived from Tuteliq output

Cooperation with these measures is a condition of continued service.

Permitted research access

This prohibition does not apply to authorised researchers with an explicit written research or dataset access agreement (TUT-100, Coded-Terms-Slang), participants in Tuteliq's formal research beta programmes operating under separate data-use agreements, or academic researchers submitting results to our held-out evaluation benchmarks under those benchmarks' terms. To apply, contact research@tuteliq.ai with your institution, research question and proposed methodology.

5. Payment and Billing

5.1 Pricing and payment terms

  • Pricing is specified in your subscription agreement
  • Billing occurs monthly or annually per your plan
  • Payment is due within the specified period, typically 30 days
  • Invoices are provided electronically

5.2 Free 14-day trial

  • Full access to plan features during the trial
  • Automatically converts to a paid plan if not cancelled
  • Cancel any time during the trial with no charge

5.3 Monthly plans

  • Billed monthly on your subscription anniversary
  • No refunds for partial months
  • Cancel any time; cancellation takes effect at the end of the billing cycle

5.4 Annual plans

  • Billed once per year in advance
  • No refunds after purchase; remaining time is not refunded on cancellation
  • Refund policy may differ for enterprise customers

5.5 Late payment

  • Late payments accrue interest at 2% per month
  • Service may be suspended if payment is 30 days late
  • Payment remains due for service during suspension

5.6 Price changes

  • Price increases are announced with 30 days' notice
  • Effective at the next billing period
  • You may terminate if you do not accept an increase

6. Warranties and Disclaimers

6.1 Limited warranties

We warrant that services are provided in a professional manner, that reasonable security measures are maintained, that we operate in line with GDPR, COPPA and KOSA obligations, that we target 99.5% uptime, and that our API documentation is accurate.

6.2 Disclaimers

Services are provided "as is". We make no warranties regarding:

  • Accuracy: detection is 97–99% accurate; false positives and negatives occur
  • Completeness: we cannot guarantee detection of all harmful content
  • Fitness: services may not be suitable for all use cases
  • Performance: response times may vary
  • Availability: interruptions may occur despite our 99.5% target
  • Merchantability: services are provided without guarantee of utility

6.3 Detection accuracy limitations

  • True positive rate: 97–99%
  • False positive rate: 1–3%
  • False negative rate: 1–3%
  • AI detection has inherent limitations and may misread sarcasm, context or cultural references
  • Emerging harmful patterns may not be detected initially

6.4 Your responsibility for decisions

Detection results are recommendations only. You must:

  • Review flagged content independently
  • Use human judgment before taking action
  • Not rely solely on our detection for compliance
  • Make your own determination of policy violations
  • Take responsibility for decisions based on our results

We are a tool to support human decision-making, not a replacement for human judgment.

7. Limitations of Liability

7.1 Liability cap

Our total liability does not exceed the greater of the amount you paid Tuteliq in the 12 months preceding the claim, or USD 100. This cap applies to all claims regardless of form.

7.2 Excluded liability

  • Indirect damages: lost profits, lost revenue, lost opportunity
  • Consequential damages: business interruption, reputational harm
  • Incidental damages: unexpected costs, workarounds
  • Special or punitive damages
  • Third-party claims arising from other users' content or actions
  • Harm arising from decisions you take on the basis of detection results

7.3 Exceptions to the cap

  • Gross negligence or wilful misconduct
  • Death or personal injury
  • Fraud or criminal conduct
  • Violation of GDPR Article 32 (security)
  • Breach of confidentiality

8. Indemnification

You agree to indemnify and hold harmless Tuteliq from:

  • Claims arising from your use of the services
  • Your violation of applicable law
  • Content you submitted
  • Your unauthorised access to others' data
  • Your use of services to harm third parties
  • Any breach of your representations and warranties

9. Dispute Resolution and Governing Law

9.1 Governing law and jurisdiction

  • These Terms are governed by Swedish law, without regard to conflict principles
  • Gävle District Court (Gävle tingsrätt) has exclusive jurisdiction as court of first instance

9.2 Process before litigation

  • Good-faith negotiation (15 days): send written notice to support@tuteliq.ai and meet to attempt resolution.
  • Escalation (15 days): escalate to management for further resolution.
  • Mediation (optional, 30 days): either party may propose mediation, costs split equally.
  • Litigation: filed with Gävle District Court as court of first instance under Swedish law.

10. Enterprise Agreements

For enterprise customers we offer custom agreements covering:

  • Custom service level agreements
  • Custom liability and indemnification terms
  • Custom payment terms and refund policies
  • Custom data processing requirements
  • Custom support and response times

Enterprise terms may differ significantly from these self-service terms. Contact sales@tuteliq.ai.

11. General Provisions

  • Entire agreement: these Terms supersede all prior negotiations
  • Amendments: we may amend these Terms; material changes require 30 days' notice, and continued use constitutes acceptance
  • Severability: invalid provisions are severed; the remainder stays in effect
  • Assignment: you may not assign without our consent; we may assign to a successor or affiliate
  • Counterparts: these Terms may be executed in counterparts

12. Contact Information

  • Legal: legal@tuteliq.ai
  • Data protection: dpo@tuteliq.ai
  • Enterprise: sales@tuteliq.ai

Document version 2.1 (zero-storage model) · Last updated August 6, 2026 · See also the Privacy Policy and DPA.

Get Started Free · Read Documentation · View Pricing