blade
Get your API key Model inference and optional tools have different paths
Blade’s own GPU infrastructure is operated by Shadow in Europe. The API also documents external-provider models that require account enablement. Confirm the selected model’s provider and hosting scope; a platform-level statement does not describe every optional route.
Flash by Staan adds web search to supported models on enabled accounts. The API reference identifies the integration as Staan WebSearch4AI and states that search queries are processed outside Shadow infrastructure. This alone does not establish a location outside Europe.
What the API reference establishes
| Topic | Documented behaviour |
|---|---|
| Web search | Requires account access and a model with native tool calling. Queries go to Staan; executed searches are separately billed. |
| External inference providers | Require the organisation’s external-provider setting to be enabled. |
| Responses conversation state | Stateless requests; store: true and previous_response_id are unsupported. |
Source: Blade API reference, checked on 16 September 2026.
Confirm the data terms for your integration
The API reference does not, by itself, establish retention periods, logging policies, training exclusions or the complete list of processing locations and subcontractors. The stateless Responses contract concerns conversation persistence; it is not a statement about every operational log.
Read the Privacy Policy, Data Processing Agreement and list of subprocessors for the applicable terms and processing scope. Confirm which model provider and optional tools your integration uses; the API contract alone does not establish a general retention or training policy.