Security & Privacy

Your manufacturing data security is our top priority. Learn how we protect your information.

  • SOC 2 Type 2
  • AI models never train on your data

If you have any questions email our security team [email protected].

Infrastructure Security

Threaded's security infrastructure is built on industry-leading providers and services:

Cloud Infrastructure

  • Amazon Web Services (AWS) - Enterprise-grade cloud hosting with SOC 2 compliance
  • Cloudflare - All traffic routes through Cloudflare's global network, providing DDoS protection, web application firewall (WAF), bot management, rate limiting, DNS security (DNSSEC), and TLS termination before requests reach our servers. Parts of our application also run on Cloudflare's edge platform — Workers, Durable Objects, KV, and R2 — keeping latency-sensitive and storage-heavy operations close to users while remaining within the same secure perimeter.
  • AWS RDS - Managed database services with automated backups and encryption

Security & Monitoring

  • AWS GuardDuty - Threat detection and continuous security monitoring
  • AWS CloudTrail - Comprehensive audit logging and compliance tracking
  • Let's Encrypt - SSL/TLS certificates with automatic renewal

Authentication & Access

  • AWS IAM - Fine-grained access controls and role-based permissions

Data & Communications

  • AWS KMS - Key management and encryption services
  • Mailgun - Secure email delivery

AI & Data Privacy

Threaded uses AI to enhance your manufacturing workflows while maintaining strict data privacy and security standards:

AI Model Usage

  • Inference Only - We use AI models strictly for inference to assist your workflows, never for training
  • Anthropic Claude - Advanced AI assistant with constitutional AI safety principles
  • OpenAI GPT Models - Industry-leading language models with enterprise data protection
  • Google Gemini - Multimodal AI model with enterprise-grade security
  • xAI Grok Models - Large language models from xAI, under the same enterprise API terms

Data Protection

  • Zero Training Data - Your prompts and data are never used to train or improve AI models
  • Enterprise API Terms - All AI providers operate under enterprise agreements with data isolation
  • Prompt Isolation - User queries are processed independently without cross-customer data exposure

How It Works

When you interact with Threaded's AI chat features, your data follows this secure process:

  1. 1
    Your request is encrypted and sent to our secure infrastructure
  2. 2
    We format your query with manufacturing context while maintaining data isolation
  3. 3
    The AI model processes your request using inference only - no data retention or training

    Inference uses pre-trained model weights to generate responses without modifying the model or storing your data

  4. 4
    The response is delivered back to you through our encrypted channels
How Threaded's AI works with your plant →

Trust Center

SOC 2 Type II audited. Request the audit report and see our security documentation in the Trust Center.

Visit the Trust Center →