Genaya AI
Ask plain-language questions about your own data.
/ai/askAsk Genaya AI about your data
Asks a plain-language question about the organization's own records and returns a written answer. Requires scope: ai:read. The model never writes queries directly. It calls a fixed set of read-only tools and Genaya executes them, always scoped to the organization behind the API key - so a question can never reach another tenant's data. What it can see is derived from the key's own read scopes, and is limited to record types the key could already read through the REST endpoints. Asking the AI can never surface data that a direct GET with the same key would refuse. A key with no read scopes at all is rejected. Each answer is 10 Genaya AI credits (a question), exactly like a question in the app's chat: from the organization's included credits, then the prepaid balance at 1 cent a credit on the plans that keep going (Pro and Business; Individual after an owner turns it on). A failed answer spends none. When the credits are spent the API answers 403 permission_error with code insufficient_balance; a trial whose member has not verified their email answers code email_unverified; a pool that could not be read answers 503 service_unavailable with Retry-After. The request weight (10) on the API request meter is separate. `tier` is accepted and ignored - one engine answers every plan.
Request body
questionrequiredtierResponse 200
objectquestionanswerblockstiermodelcreditscollectionsQueriedtoolCallsUsed