Class Cortex

constructor

• new Cortex(ops): common.apis.cortex.Cortex

• Parameters

  • ops: common.apis.cortex.CortexOps

  Returns common.apis.cortex.Cortex

CortexRAM

apiBaseUrl

function_dictionary

function_input_ch

functions

Optionalinsights

is_running_function

last_result

OptionalllmCallFn

log

messages

model

name

Optionaloutput_structure

OptionalprocessManager

promptManager

prompt_history

provider

runner

sandbox

Optionalscm

system_msg

usage

user_output

utilities

Optionalworkspace

_cortex_msg

• _cortex_msg(content): common.apis.cortex.CortexMessage

• Parameters

  • content: string

  Returns common.apis.cortex.CortexMessage

_user_msg

• _user_msg(content): common.apis.cortex.UserMessage

• Parameters

  • content: string

  Returns common.apis.cortex.UserMessage

add_cortex_message

• add_cortex_message(msg): void

• Add CortexMessage

  Parameters

  • msg: common.apis.cortex.CortexMessage

  Returns void

add_cortex_output

• add_cortex_output(output): void

• Parameters

  • output: common.apis.cortex.CortexOutput

  Returns void

add_user_data_input

• add_user_data_input(data, type?): void

• Parameters

  • data: any
  • type: string = 'code_result'

  Returns void

add_user_message

• add_user_message(msg): void

• Add UserMessage

  Parameters

  • msg: common.apis.cortex.UserMessage

  Returns void

add_user_result_input

• add_user_result_input(codeExecutionResult): void

• Parameters

  • codeExecutionResult: common.apis.cortex.CodeExecutionResult

  Returns void

  Deprecated

  Use add_user_data_input

add_user_text_input

• add_user_text_input(text): void

• Parameters

  • text: string

  Returns void

build_messages

• build_messages(): {
      content: string;
      role: string;
  }[]

• Build the message array

  Returns {
      content: string;
      role: string;
  }[]

build_sandbox_context

• build_sandbox_context(): Record<string, any>

• Builds sandbox context with all cortex functions

  Returns Record<string, any>

cancel

• cancel(): void

• Cancel any in-flight LLM run. The runner will throw CortexCancelledError.

  Returns void

chat

• chat(text, maxLoops?): Promise<string>

• Convenience method: Send a message and run LLM in one call

  Parameters

  • text: string

    The user's message
  • maxLoops: number = 4

    Maximum function calling loops (default: 4)

  Returns Promise<string>

  Promise - The LLM response

collect_args

• collect_args(arg_array): any

• Parameters

  • arg_array: string[]

  Returns any

configure_user_output

• configure_user_output(fn): void

• Parameters

  • fn: any

  Returns void

createRunnerContext

• createRunnerContext(signal?): common.apis.cortex.RunnerContext

• Create a RunnerContext that wraps this Cortex instance's state and methods. The runner uses this to interact with Cortex without direct coupling.

  Parameters

  • Optionalsignal: AbortSignal

  Returns common.apis.cortex.RunnerContext

emit_event

• emit_event(evt): void

• Parameters

  • evt: any

  Returns void

getContextStatus

• getContextStatus(): common.apis.cortex.ContextStatus

• Get current context status including token usage breakdown

  Returns common.apis.cortex.ContextStatus

getStateSnapshot

• getStateSnapshot(): {
      cortexRAM: Record<string, any>;
      last_result: any;
      workspace: Record<string, any>;
  }

• Returns {
      cortexRAM: Record<string, any>;
      last_result: any;
      workspace: Record<string, any>;
  }

  • cortexRAM: Record<string, any>

  • last_result: any

  • workspace: Record<string, any>

getUsage

• getUsage(): common.apis.cortex.UsageStats

• Get current usage stats

  Returns common.apis.cortex.UsageStats

get_var

• get_var(id): any

• Parameters

  • id: string

  Returns any

handle_function_call

• handle_function_call(fCall): Promise<{
      error: any;
      name: string;
      result: any;
  }>

• Parameters

  • fCall: common.apis.cortex.FunctionCall

  Returns Promise<{
      error: any;
      name: string;
      result: any;
  }>

handle_function_input

• handle_function_input(i): Promise<void>

• Parameters

  • i: any

  Returns Promise<void>

handle_llm_response

• handle_llm_response(fetchResponseOrData, loop, fetchTiming?): Promise<any>

• Parameters

  • fetchResponseOrData: any
  • loop: number
  • OptionalfetchTiming: {
        elapsed: number;
        end: number;
        start: number;
    }

    • elapsed: number

    • end: number

    • start: number

  Returns Promise<any>

initProcessManager

• initProcessManager(sandboxFactory): ProcessManager

• Parameters

  • sandboxFactory: common.apis.cortex.SandboxFactory

  Returns ProcessManager

log_event

• log_event(msg): void

• Parameters

  • msg: string

  Returns void

rerun_llm_with_output_format

• rerun_llm_with_output_format<T>(options): Promise<TypeOf<T>>

• Re-run LLM with a different output format using existing conversation history

  This "time travels" by:

  1. Taking the current message history (minus the last cortex message that triggered this call)
  2. Rebuilding the system prompt with section overrides via PromptManager
  3. Running a structured completion with the custom schema

  By default, preserves all original prompt sections. Use sectionOverrides to:

  • Replace section args: { responseGuidance: ['Custom guidance...'] }
  • Exclude a section entirely: { functions: null }

  Useful when a function needs the LLM to extract structured data from the conversation but the main CortexOutput format isn't granular enough.

  Type Parameters

  • T extends ZodType<any, ZodTypeDef, any>

  Parameters

  • options: {
        schema: T;
        schema_name: string;
        sectionOverrides?: common.apis.cortex.SectionOverrides;
    }

    • schema: T

    • schema_name: string

    • OptionalsectionOverrides?: common.apis.cortex.SectionOverrides

  Returns Promise<TypeOf<T>>

resetUsage

• resetUsage(): void

• Reset usage stats

  Returns void

resolve_args

• resolve_args(args): any

• Parameters

  • args: any

  Returns any

resolve_cortex_ram_reference

• resolve_cortex_ram_reference(v): any

• Parameters

  • v: string

  Returns any

run_code_output

• run_code_output(output): Promise<common.apis.cortex.FunctionResult>

• Executes JavaScript code in sandbox instead of running function calls

  Parameters

  • output: common.apis.cortex.CodeOutput

  Returns Promise<common.apis.cortex.FunctionResult>

run_cortex_output

• run_cortex_output(co): Promise<common.apis.cortex.FunctionResult>

• Parameters

  • co: common.apis.cortex.CortexOutput

  Returns Promise<common.apis.cortex.FunctionResult>

run_llm

• run_llm(loop?): Promise<string>

• Run the LLM with the specified system message, message history, and parameters If loop=N, after obtaining a function response another LLM call will be made automatically, until no more functions are called or until N calls have been made

  Delegates to the current runner. Kept for backward compatibility.

  Parameters

  • loop: number = 6

  Returns Promise<string>

run_structured_completion

• run_structured_completion<T>(options): Promise<TypeOf<T>>

• Run a structured completion with a custom Zod schema Allows functions to invoke their own LLM completions with custom output formats

  Type Parameters

  • T extends ZodType<any, ZodTypeDef, any>

  Parameters

  • options: {
        messages: {
            content: string;
            role: "user" | "assistant" | "system";
        }[];
        schema: T;
        schema_name: string;
    }

    • messages: {
          content: string;
          role: "user" | "assistant" | "system";
      }[]

    • schema: T

    • schema_name: string

  Returns Promise<TypeOf<T>>

setRunner

• setRunner(runner): void

• Swap the runner and rebuild the system prompt with the runner's prompt format.

  Parameters

  • runner: common.apis.cortex.Runner

  Returns void

set_is_running_function

• set_is_running_function(v): void

• Running function

  Parameters

  • v: boolean

  Returns void

set_var

• set_var(v): Promise<string>

• Parameters

  • v: any

  Returns Promise<string>

set_var_with_id

• set_var_with_id(v, id): Promise<string>

• Parameters

  • v: any
  • id: string

  Returns Promise<string>

updateUsage

• updateUsage(promptTokens, completionTokens, cachedInputTokens?): void

• Update usage stats after an LLM call

  Parameters

  • promptTokens: number
  • completionTokens: number
  • cachedInputTokens: number = 0

  Returns void