Sub-Processors

The third parties who process data on the Cognethics platform. Every add, edit, and retirement is published the moment it lands in the registry.

Current sub-processors

Sub-processor Role Purpose Hosting location Data categories Contract BAA
Amazon Web Services Infrastructure Compute, storage, database, networking, and managed security (CloudTrail, KMS, GuardDuty, Security Hub, VPC Flow Logs). Sensitive application data is encrypted with a dedicated per-tenant AWS KMS customer-managed key (CMK). AWS region and availability zone selected by the customer (any AWS region), with a cross-region hot standby all customer data (encrypted at rest); operational logs; encryption key material (per-tenant customer-managed keys); infrastructure telemetry DPA + BAA Signed
Anthropic LLM Provider Large language model inference for AI features (summaries, classifications, generation, agent reasoning). Accessed via Anthropic API directly and via AWS Bedrock. Anthropic (US) and AWS Bedrock (US regions) user prompts; document content for AI processing Data Processing Agreement Signed
Google (Gemini API) LLM Provider Large language model inference for specific workflows including healthcare EOB extraction, denial analysis, technical specification extraction, and invoice parsing. Google (US regions) document content for AI processing; extracted structured data Data Processing Agreement
Cloudflare CDN / Edge Edge CDN, DDoS protection, and Cloudflare Tunnel routing for *.cognethics.com. Global edge network request metadata (IP, timestamp, User-Agent, referrer) Data Processing Agreement

Data processing details

Amazon Web Services (AWS)

Region: AWS region and availability zone selected by the customer (any AWS region), with a cross-region hot standby for regional disaster recovery

Services used:

BAA status: Signed via AWS Artifact

Encryption: All data encrypted at rest via AWS KMS, with a dedicated per-tenant customer-managed key (CMK) for sensitive application data, and in transit (TLS 1.2+)

Compliance: HIPAA-compatible infrastructure; subject to AWS BAA

Anthropic

Service: Large language models (LLM)

Use case: Primary LLM for AI features (summaries, classifications, generation, agent reasoning)

Data sent: User prompts, document content (encrypted in transit)

Access pattern: Anthropic API directly and via AWS Bedrock

Contract: Anthropic commercial terms; Bedrock traffic covered under the AWS BAA

Google (Gemini API)

Service: Large language models (LLM) — Gemini family

Use case: Narrow extraction workflows — healthcare EOB extraction and denial analysis, spare-parts technical specification extraction, invoice parsing

Data sent: Document content for extraction, structured output (encrypted in transit)

Contract: Google Cloud Data Processing Addendum

Cloudflare

Service: CDN, DDoS protection, request routing

Data sent: Request metadata only (IP address, timestamp, User-Agent, referrer)

No access to: Customer data, documents, database content

Purpose: Performance, security, availability

How we notify you of changes

The sub-processor list is a live registry — every add, edit, and retirement is published the moment it lands in the registry.

Every change carries a change_log_entry describing what changed and why, so subscribers can review impact without reading code.

Requesting more information

Pick the matching topic below and we'll route your request to the right team:

Request sub-processor information

Pick a topic and we'll route your message to the right team.

Related policies

← Back to Trust Center