We do not mix model providers. Every piece of model work in Asedra runs on Claude by Anthropic, through your company's own Anthropic account. Permissions and actions are enforced by the system, not by the model. Here is why we chose one model family, and what Anthropic itself publishes.
Package evaluations and release checks target one model family. A test result applies to what actually runs for you, not to one of several routes.
Your compliance and security teams read one provider's terms instead of comparing several.
Your company contracts with Anthropic directly. Usage, cost and rate limits are visible in the Claude Console by model and by key.
Anthropic publishes Claude's constitution and its Responsible Scaling Policy, so you can read them before you choose, not after.
Claude Code handles anything from a bug fix to a multi-day migration. The Agent SDK gives production systems the same tools, work loop and context management, and runs non-interactively. That is the machinery that can hold a long task end to end: from gathering context to a checked, prepared answer.
The catch is that it is built for engineers. Asedra is the layer that lets a salesperson, a controller or an HR partner benefit from it without ever touching a terminal.
| Topic | What Anthropic states |
|---|---|
| Training | By default, inputs and outputs from commercial products, including the API, are not used to train models; exceptions are feedback the customer explicitly submits or another explicit opt-in. |
| Ownership | Under the Commercial Terms, the customer retains rights to its inputs and owns its outputs, as between the parties and to the extent permitted by law. |
| Certifications | ISO 27001:2022, ISO/IEC 42001:2023, SOC 2 Type I and Type II for Anthropic's commercial products. |
These are Anthropic's terms and certifications, not Asedra's. Asedra has no certifications of its own yet and displays none. Sources and check dates are listed in the architecture overview.
We are not an Anthropic partner, and Anthropic has not endorsed or certified Asedra. We simply build only on Claude.
No Anthropic or Claude logos appear on this site, and neither name is part of our product name.
We send a written description of which data goes where.