Skip to main content

context.history_archive

History archive — persists the full conversation history for search.

The full conversation history (before compression) is continuously saved to a JSON file. A search_conversation_history tool allows the agent to look up specific details from compressed messages.

HistoryArchive Objects​

class HistoryArchive()

Persist and search the full conversation history.

Parameters​

archive_path : str Path to the JSONL archive file.

append​

def append(message: dict[str, Any]) -> None

Append a message to the archive.

append_many​

def append_many(messages: list[dict[str, Any]]) -> None

Append multiple messages to the archive.

def search(query: str, max_results: int = 10) -> list[dict[str, Any]]

Search the archive for messages matching the query.

Uses simple case-insensitive substring matching.

Parameters​

query : str Search text. max_results : int Maximum number of results.

Returns​

list[dict] Matching archived messages (most recent first).

get_total_messages​

def get_total_messages() -> int

Return the total number of archived messages.

search_conversation_history_tool_fn​

def search_conversation_history_tool_fn(
        archive: HistoryArchive) -> dict[str, Any]

Create a tool definition for the agent to search conversation history.

Returns a dict with name, description, parameters and function suitable for registration as a PydanticAI tool.