LLM servers, providers, and agent Nodes

API reference for the LLM servers, providers, and agents library nodes in chia.models.

Base class

class chia.base.llm_call.QueryResult(result: str, returncode: int, stderr: str, stream_result: str, success: bool = False)[source]

Bases: object

Structured result from prompting an LLM or agent.

Parameters:

result (str) – The final response from the LLM or agent
returncode (int) – The returncode from running the prompt
stderr (str) – The stderr output from running the prompt (clis only)
stream_result (str) – The full transcript of all turns of the LLM or agent
success (bool) – Whether the prompt completed successfully

class chia.base.llm_call.LLMCallBase(system_message: str)[source]

Bases: ABC

Polymorphic base container for generic LLM and agent configuration traits and behavior. Easy to switch between different backing providers, servers, and CLIs

abstractmethod prompt(user_message: str, tools: List[ChiaTool] | None = []) → QueryResult[source]

Send a prompt to this LLM

Parameters:

user_message (str) – Message used to prompt the LLM
tools (Optional[List[ChiaTool]]) – Tools available to the LLM during the call

Providers

We provide the prompt interface for each of the following providers. Many of our interfaces for these providers are lightly tested using free tokens. In particular, error handling is likely fragile.

LLM servers, providers, and agent Nodes

Base class

Providers

CLIs

Remote providers

On-prem servers