Skip to main content

Class: LLM

Extends​

  • BaseClient

Constructors​

Constructor​

new LLM(ctx?, _id?, _historyJSON?, _lastReply?, _model?, _provider?, _sync?, _tools?): LLM

Constructor is used for internal usage only, do not create object from it.

Parameters​

ctx?​

Context

_id?​

LLMID

_historyJSON?​

string

_lastReply?​

string

_model?​

string

_provider?​

string

_sync?​

LLMID

_tools?​

string

Returns​

LLM

Overrides​

BaseClient.constructor

Methods​

attempt()​

attempt(number_): LLM

create a branch in the LLM's history

Parameters​

number_​

number

Returns​

LLM

bindResult()​

bindResult(name): Binding

returns the type of the current state

Parameters​

name​

string

Returns​

Binding

env()​

env(): Env

return the LLM's current environment

Returns​

Env

history()​

history(): Promise<string[]>

return the llm message history

Returns​

Promise<string[]>

historyJSON()​

historyJSON(): Promise<string>

return the raw llm message history as json

Returns​

Promise<string>

id()​

id(): Promise<LLMID>

A unique identifier for this LLM.

Returns​

Promise<LLMID>

lastReply()​

lastReply(): Promise<string>

return the last llm reply from the history

Returns​

Promise<string>

loop()​

loop(): LLM

synchronize LLM state

Returns​

LLM

model()​

model(): Promise<string>

return the model used by the llm

Returns​

Promise<string>

provider()​

provider(): Promise<string>

return the provider used by the llm

Returns​

Promise<string>

sync()​

sync(): Promise<LLM>

synchronize LLM state

Returns​

Promise<LLM>

tokenUsage()​

tokenUsage(): LLMTokenUsage

returns the token usage of the current state

Returns​

LLMTokenUsage

tools()​

tools(): Promise<string>

print documentation for available tools

Returns​

Promise<string>

with()​

with(arg): LLM

Call the provided function with current LLM.

This is useful for reusability and readability by not breaking the calling chain.

Parameters​

arg​

(param) => LLM

Returns​

LLM

withEnv()​

withEnv(env): LLM

allow the LLM to interact with an environment via MCP

Parameters​

env​

Env

Returns​

LLM

withModel()​

withModel(model): LLM

swap out the llm model

Parameters​

model​

string

The model to use

Returns​

LLM

withPrompt()​

withPrompt(prompt): LLM

append a prompt to the llm context

Parameters​

prompt​

string

The prompt to send

Returns​

LLM

withPromptFile()​

withPromptFile(file): LLM

append the contents of a file to the llm context

Parameters​

file​

File

The file to read the prompt from

Returns​

LLM

withQuery()​

withQuery(): LLM

Provide the entire Query object to the LLM

Returns​

LLM

withSystemPrompt()​

withSystemPrompt(prompt): LLM

Add a system prompt to the LLM's environment

Parameters​

prompt​

string

The system prompt to send

Returns​

LLM