Skip to main content
POST
llm.complete

Authorizations

Authorization
string
header
required

API key: sk_live_…

Headers

Idempotency-Key
string

Optional. A retry with the same key returns the stored result.

Maximum string length: 255
Pramaan-Max-Price-Micro
string

Optional. The most this request may be priced at, in micro-USD (a whole number above 0). It replaces the capability's own price ceiling for this request, lower or higher. Sources that could cost more are left out; when none is left the answer is 400 with the lowest limit that would work.

Pattern: ^[1-9]\d{0,14}$

Body

application/json
messages
object[]
required

The chat so far, oldest first: 1 to 500 messages, at most 1,000,000 characters in total, with at least one user or assistant message.

Required array length: 1 - 500 elements
model
enum<string>
default:balanced

Quality tier: "fast" is cheapest, "best" is most capable. Default "balanced".

Available options:
fast,
balanced,
best
maxTokens
number
default:4096

Most tokens to generate, from 1 to 1048576; each source enforces its model maximum. Default 4096. The cap can include internal reasoning; very small values may produce no text.

Required range: 1 <= x <= 1048576
jsonSchema
object

JSON Schema subset: type (including nullable arrays), properties, required, additionalProperties, items, enum, const, anyOf, acyclic local $defs/$ref, minimum/maximum/exclusiveMinimum/exclusiveMaximum, minLength/maxLength, minItems/maxItems, title/description/default/examples. $schema, $id and $comment are accepted and ignored. format is accepted and ignored; the model is not told about it and results are not checked against it. Any other keyword is refused, and the error names it and where it is. Maximum 64 KiB, 256 nodes, depth 16; no external or recursive references. JSON text is preserved exactly; invalid or oversized output fails.

temperature
number

Randomness from 0 (steady) to 2 (varied). Provider default when omitted.

Required range: 0 <= x <= 2

Response

The result

text
string
required

The generated reply.

finishReason
enum<string>
required

Why it stopped: "stop" (finished), "length" (hit maxTokens) or "other".

Available options:
stop,
length,
other
usage
object
required

Token counts for the call.