How we use AI

Agent tasks, code reviews, chat, and estimates are powered by third-party AI providers: Anthropic (Claude models), open models routed through OpenRouter, and DigitalOcean serverless inference. We name providers rather than specific model versions because the models in use are configured per deployment and improve over time.

Requests are routed to the best available provider for the task and your plan. If a provider is unavailable or rate-limited, the platform automatically falls back to the next configured provider, so your task keeps running.

When you start a task, your prompt and the repository context relevant to that task — the files an agent needs to read or change — are sent to the provider handling the request. Data is only sent for tasks you initiate; connecting a repository does not by itself send anything to an AI provider.

Code produced by Factory Nexus agents for your projects is yours. We claim no ownership over your prompts, repositories, or generated output.

Factory Nexus does not use your code or prompts to train models. Your content is processed to complete the tasks you run — nothing more.