package ocaml-ai-sdk

  1. Overview
  2. Docs
OCaml AI SDK - Provider abstraction for AI models

Install

dune-project
 Dependency

Authors

Maintainers

Sources

ocaml-ai-sdk-0.6.tbz
sha256=44e7f6ebef6da2828ec2844dd5b38c8f37284c08e54a29e06f6a8822df6f8e10
sha512=ce31d0868abf28bcaef7b24a0ab09b37dcfc94db3de96b38ff36a03800ae59a19e8009a35314f93191b3c8f178c40f8569630d4389f56ddc2f2cc94fb7bffc3c

doc/ocaml-ai-sdk.ai_core/Ai_core/Generate_text/index.html

Module Ai_core.Generate_textSource

Non-streaming text generation with multi-step tool execution.

Calls the provider model, executes tools if requested, feeds results back, and loops until the model stops or max_steps is reached.

Each provider call is retried on retryable errors and transient network failures with exponential backoff (see Retry.with_retries).

  • parameter max_retries

    Number of additional attempts per provider call (default 2). Set to 0 to disable retries.

  • parameter max_retry_delay_ms

    Caps backoff; a server Retry-After above it stops retries instead of sleeping. Uncapped by default.

Sourceval generate_text : model:Ai_provider.Language_model.t -> ?system:string -> ?system_provider_options:Ai_provider.Provider_options.t -> ?prompt:string -> ?messages:Ai_provider.Prompt.message list -> ?tools:(string * Core_tool.t) list -> ?tool_choice:Ai_provider.Tool_choice.t -> ?output:(Yojson.Basic.t, Yojson.Basic.t) Output.t -> ?max_steps:int -> ?max_retries:int -> ?max_retry_delay_ms:int -> ?stop_when:Stop_condition.t list -> ?max_output_tokens:int -> ?temperature:float -> ?top_p:float -> ?top_k:int -> ?stop_sequences:string list -> ?seed:int -> ?headers:(string * string) list -> ?provider_options:Ai_provider.Provider_options.t -> ?on_step_finish:(Generate_text_result.step -> unit) -> ?telemetry:Telemetry.t -> ?pending_tool_approvals:Generate_text_result.pending_tool_approval list -> unit -> Generate_text_result.t Lwt.t