Trust and Transparency

Authorized Sub-processors

Effective Date: May 15, 2026 — Last Updated: May 15, 2026


Overview

DEVUP AI utilizes a select number of industry-standard infrastructure partners ("sub-processors") to deliver the Platform securely and reliably. Each sub-processor is contractually bound by a Data Processing Agreement (DPA) that governs data access, retention, and security obligations in accordance with Algerian Law 18-07 on the protection of personal data.

This page provides a transparent disclosure of the sub-processors currently authorized to process data on behalf of DEVUP AI. For questions about our sub-processor practices, contact us at support@devupai.com.

Authorized Sub-processor List

Sub-processorPurposeData ProcessedData Center RegionRetention Policy
SupabaseDatabase hosting, user authentication, and session managementAccount data (email, hashed credentials), billing records, API usage metadata (timestamps, token counts, costs). Does NOT include inference content (prompts/completions).United States (US East)As defined in our Privacy Policy — account data retained for account lifetime + 30 days; billing records retained for 10 years per Algerian tax law.
VercelFrontend hosting, edge functions, and API gateway routingTechnical data (IP address, request headers, browser metadata) for edge routing and serverless function execution. Does NOT include inference content.Global Edge NetworkTransient — request-level data is processed in real-time for routing and not persisted beyond standard serverless execution logs.
Tier-1 Cloud GPU Infrastructure
(Nebius, RunPod, Lambda Labs)
Ephemeral AI model inference and compute. We leverage a distributed network of leading global compute providers to execute AI inference workloads. Our infrastructure network includes, but is not limited to, Nebius, RunPod, and Lambda Labs.Inference payloads (prompts and completions) transmitted securely for immediate processing. Held strictly in volatile memory (RAM/VRAM).International (distributed across provider regions)Zero persistent storage. Payloads are transmitted for immediate processing, held strictly in volatile memory (RAM), and instantly purged. None of these sub-processors are permitted to store customer inference data or use it for model training.

Infrastructure Compliance & Certifications

PartnerInfrastructure RoleData Processing LimitsCertifications
Nebius Logo
Nebius
Ephemeral AI ComputeVolatile memory only. Instant purge. No training.
SOC 2 Type IIISO 27001GDPR
RunPod Logo
RunPod
Ephemeral AI ComputeVolatile memory only. Instant purge. No training.
SOC 2 Type IIISO 27001GDPR
Lambda Labs Logo
Lambda Labs
Ephemeral AI ComputeVolatile memory only. Instant purge. No training.
SOC 2 Type IIISO 27001GDPR
Supabase Logo
Supabase
Database & AuthenticationAccount & billing data only. No inference content.
SOC 2 Type IIISO 27001GDPR
Vercel Logo
Vercel
Frontend Hosting & Edge RoutingTransient routing metadata. No persistent inference logs.
SOC 2 Type IIISO 27001GDPR

The compliance certifications listed above are maintained independently by our respective Tier-1 infrastructure sub-processors. DEVUP AI leverages these highly secure environments to ensure professional data protection.

Data Processing Safeguards

All authorized sub-processors are subject to the following requirements:

Changes to This List

DEVUP AI reserves the right to update this sub-processor list as infrastructure needs evolve. Material changes (such as adding a new category of sub-processor or changing the data center region for persistent data) will be reflected on this page with an updated "Last Updated" date. We encourage users to review this page periodically.


Contact

For questions regarding our sub-processor practices or to request a copy of any applicable Data Processing Agreement, please contact:

For related legal documents, see our Terms of Service and Privacy Policy.