Trust Centre
Rigorous data protection, transparent provenance, and secure infrastructure. Explore our policies and security architecture ensuring your AI data pipeline meets enterprise standards.
Policies & Compliance Documentation
We provide full transparency into our operations to streamline enterprise procurement and legal review.
Data Provenance Policy
Proves "no scraping" and details our rigorous legal sourcing framework for all datasets.
[DOWNLOAD POLICY]Consent Policy
Critical documentation outlining verifiable participant consent for speech, biometric, image, and medical data.
[DOWNLOAD POLICY]Dataset Licensing Model
Clear terms on commercial use, training rights, resale limitations, and exclusivity agreements.
[VIEW TERMS]DPA & SCC Availability
Data Processing Agreements and Standard Contractual Clauses ready for EU enterprise buyers.
[REQUEST DPA]Retention & Deletion
Granular policies detailing secure data lifecycle management, necessary for handling sensitive client data.
[DOWNLOAD POLICY]AI Ethics Policy
Our commitment to fairness, bias mitigation, and ethical representation in model training.
[VIEW POLICY]Sample Data Card
A transparent schema detailing exactly how our datasets are structured, tagged, and measured.
[VIEW SAMPLE]Security & Privacy Controls
SYS.COMPLIANTSOC 2 Aligned
Our security measures align with SOC 2 principles for security, availability, and processing integrity.
ISO 27001 Aligned
Information security management controls modeled on internationally recognized standards.
GDPR & CCPA
Strict compliance with European and California data privacy and consent regulations.
Zero-Trust Data Pipelines
Our annotation and delivery infrastructure is built on zero-trust principles. We separate client storage environments from public networks, applying least-privilege access across the entire ML lifecycle.
- AES-256 Encryption Data encrypted at rest and in transit via TLS 1.3.
- Isolated VPCs Dedicated client endpoints restricted to authorized IPs.
- Audit Logging Immutable interaction logs for comprehensive pipeline visibility.
{
"pipeline": {
"encryption": "AES-256-GCM",
"transit": "TLS 1.3",
"accessControl": "RBAC",
"mfaEnforced": true
},
"storage": {
"isolation": "Tenant-Level",
"geoPinning": true,
"retention": "Configurable"
},
"auditing": {
"frequency": "Real-time",
"retention": "365 days"
}
}
Annotator Security Model
Technology means nothing without secure operations. We manage our global annotator workforce through strict protocols, demonstrating exactly how our distributed workforce is controlled to ensure sensitive data never leaves secure environments.
Identity Verification
Biometric identity checks and continuous monitoring to verify the active annotator matches their assigned secure profile.
Secure Workspaces
Screen-recording prevention, disablement of local downloads, copy-paste restrictions, and enforced secure VPN routing.
Strict NDAs & Training
All staff execute binding confidentiality agreements and undergo mandatory compliance and security awareness training.
Procurement FAQ
What is dataset provenance?
Dataset provenance is the documented origin, chain of custody, and licensing history of training data. It proves that the data was legally sourced, avoiding unauthorized scraping and ensuring full enterprise compliance.
What is a data card?
A data card is a standardized transparency document that details a dataset's composition, collection methodology, intended use cases, and potential biases, allowing engineers to evaluate its suitability for AI training.
Do you scrape data from the public internet?
No. We maintain a strict "no scraping" policy. All datasets are legally sourced, either crowdsourced directly from consented participants, gathered from partners with full commercial rights, or synthetically generated.
Who owns the IP of the models trained on your data?
You do. Our dataset licensing model grants full rights to use the data for training, and you retain 100% intellectual property ownership of the resulting model weights and outputs.
How do you handle participant consent?
We collect verifiable, opt-in consent for all human-generated data (speech, biometric, text). Our consent frameworks comply with GDPR and CCPA requirements, ensuring participants understand exactly how their data will be used for AI training.
Can you sign a DPA or SCCs?
Yes. For EU enterprise buyers, we have standard Data Processing Agreements and Standard Contractual Clauses ready for review by your legal team.
Require Further Documentation?
We provide full transparency to our enterprise clients. Request our compliance matrix, DPAs, or speak with our security engineering team.