Reasoning & Depth

How AI Query decides how much to say, and why a complete answer is never cut short.

#How depth works

AI Query always answers from a fully computed birth chart, and the length of the response adapts to the question you ask. A narrow question — “What is my Moon sign?” — gets a short, direct answer. A multi-factor question — “How do my current dashas affect my career and relationships this year?” — gets a longer synthesis that reasons across more of the chart. Depth is not a slider you set directly; it falls out of two things: the delivery tier you call, and, on the Standard tier, which version you pin.

#The no-truncation guarantee

Note
No-truncation is a policy, not a request parameter. Short questions get short answers and cost less; longer questions get longer answers; and however long an answer runs, it is never cut off mid-thought. There is nothing to configure — the guarantee applies to every call, on every tier, by default.

#Depth by tier and version

The speed parameter selects the delivery tier — Eco, Standard, Swift or Pro Ultra — which is the main lever on both depth and latency. On the Standard tier only, the version field pins the reasoning path: "2" (the default) or "2.5", which selects Vedika Standard 2.5, a newer Standard variant. See the full tier and version comparison on the Model Catalog.

{
  "question": "What does my chart say about career timing?",
  "birthDetails": { "...": "..." },
  "speed": "standard",
  "version": "2.5"
}

Depth (how much the answer reasons about) is a separate control from shape (how the answer is formatted: text default, markdown or json) — see Structured Outputs for the responseFormat field. And regardless of depth or tier, every reading stays consistent with the computed chart — see Grounding.

Model CatalogThe four delivery tiers, compared.Structured OutputsControl response shape with responseFormat.GroundingWhy every reading stays consistent with the chart.
PreviousChoosing a ModelNextStreaming Responses