analyze_recorded_traffic

Analyze saved network traffic recordings to identify patterns, filter by criteria like host or status, and generate structured summaries for debugging and monitoring.

Instructions

Analyze a saved recording snapshot with compact summaries. Returns structured TrafficSummary items with matched_fields and match_reasons. Use get_traffic_entry_detail to drill down into a specific entry_id afterwards.

Input Schema

TableJSON Schema

Name	Required	Default
`recording_path`	No
`preset`	No	api_focus
`host_contains`	No
`path_contains`	No
`method_in`	No
`status_in`	No
`resource_class_in`	No
`min_priority_score`	No
`request_header_name`	No
`request_header_value_contains`	No
`response_header_name`	No
`response_header_value_contains`	No
`request_content_type`	No
`response_content_type`	No
`request_body_contains`	No
`response_body_contains`	No
`request_json_query`	No
`response_json_query`	No
`include_body_preview`	No
`max_items`	No
`max_preview_chars`	No
`max_headers_per_side`	No
`scan_limit`	No

Implementation Reference

charles_mcp/services/traffic_query_orchestrator.py:52-76 (handler)

The core logic for analyze_recorded_traffic, which handles orchestration of history snapshots and building the query result.

async def analyze_recorded_traffic(
    self,
    *,
    recording_path: str | None,
    query: TrafficQuery,
) -> TrafficQueryResult:
    try:
        prepared = await self.prepare_capture(
            source="history",
            query=query,
            recording_path=recording_path,
            advance=False,
        )
    except FileNotFoundError:
        return TrafficQueryResult(
            source="history",
            items=[],
            total_items=0,
            scanned_count=0,
            matched_count=0,
            filtered_out_count=0,
            filtered_out_by_class={},
            warnings=["no_saved_recordings"],
        )
    return self.build_query_result(prepared=prepared, query=query, include_items=True)

charles_mcp/tools/history.py:36-64 (handler)

The tool registration and function handler for analyze_recorded_traffic within the MCP tool definitions.

async def analyze_recorded_traffic(
    ctx: ToolContext,
    recording_path: Optional[str] = None,
    preset: TrafficPreset = "api_focus",
    host_contains: Optional[str] = None,
    path_contains: Optional[str] = None,
    method_in: Optional[list[str]] = None,
    status_in: Optional[list[int]] = None,
    resource_class_in: Optional[list[str]] = None,
    min_priority_score: Optional[int] = None,
    request_header_name: Optional[str] = None,
    request_header_value_contains: Optional[str] = None,
    response_header_name: Optional[str] = None,
    response_header_value_contains: Optional[str] = None,
    request_content_type: Optional[str] = None,
    response_content_type: Optional[str] = None,
    request_body_contains: Optional[str] = None,
    response_body_contains: Optional[str] = None,
    request_json_query: Optional[str] = None,
    response_json_query: Optional[str] = None,
    include_body_preview: bool = True,
    max_items: int = 10,
    max_preview_chars: int = 128,
    max_headers_per_side: int = 6,
    scan_limit: int = 500,
) -> TrafficQueryResult:
    """Analyze a saved recording snapshot with compact summaries.
    Returns structured TrafficSummary items with matched_fields and match_reasons.
    Use get_traffic_entry_detail to drill down into a specific entry_id afterwards."""

Charles MCP Server