AI model providers
ArchNGN uses AI model partners to perform specific parts of its workflows, including workstream and connector jobs. We encourage you to review the terms and policies of each provider before using the platform.
Current Providers
Section titled “Current Providers”The AI provider serving your account depends on the cloud environment hosting it. Your inputs are processed within the same cloud your account’s data lives in:
- AWS-hosted environments run AI workloads on Amazon Bedrock, with Anthropic Claude as the primary model family.
- Azure-hosted environments run AI workloads on Azure OpenAI, with OpenAI GPT models.
- Google Cloud-hosted environments run AI workloads on Vertex AI, with Google Gemini models.
Inputs you submit through workstreams and connectors are processed by the provider in the region configured for your environment.
You don’t choose a model. ArchNGN runs each kind of work at a fixed capability level, and your hosting environment’s provider supplies the model for that level. Designs, design reviews and policy drafting run on the provider’s most capable model; connector runs and design checks run on a standard model; indexing and search use a lightweight one.
| Provider | Service | Terms of Service | Usage Policy |
|---|---|---|---|
| Amazon Web Services | Bedrock | Service Terms | Acceptable Use Policy |
| Anthropic | Claude | Commercial Terms | Acceptable Use Policy |
| Microsoft | Azure OpenAI | Product Terms | Code of Conduct |
| Vertex AI | Service Terms | Acceptable Use Policy |
ArchNGN does not use your inputs or outputs to train foundation models. Model providers may process requests on ArchNGN’s behalf under the terms above.