search_calls

Find calls by date range, agent name, Compass-score range, or agent-lift band. Returns summaries (not full transcripts). Use get_call_detail for deep dives. Each result includes session_id, agent, date, duration, Compass score, lift, signal count, and a brief summary.

Parameters

Parameter	Type	Required	Default	Description
date_from	string	no	—	Start date (ISO format, e.g. '2026-01-01'). Optional.
date_to	string	no	—	End date (ISO format). Optional.
agent_name	string	no	—	Filter by agent name (exact match). Optional.
min_score	number	no	—	Minimum Compass Score (1.0-5.0). Optional.
max_score	number	no	—	Maximum Compass Score (1.0-5.0). Optional.
lift_band	string (enum: strong_positive, positive, neutral, negative, strong_negative)	no	—	Filter by agent lift band.
sort	string (enum: score_asc, score_desc, date_desc, date_asc, lift_desc)	no	date_desc	Sort order. Default: date_desc.
limit	integer	no	10	Max results (1-50). Default: 10.

Example call

from fastmcp import Client
from fastmcp.client.auth import BearerAuth

async with Client(
    "https://mcp.chordia.ai/mcp",
    auth=BearerAuth("YOUR_APPLICATION_KEY"),
) as client:
    result = await client.call_tool(
        "search_calls",
        {
            "date_from": "2026-01-01",
            "date_to": "2026-01-01"
        },
    )
    print(result.content[0].text)
# Plus any of the 8 optional params from the table above.

In Claude Desktop, Cursor, or any MCP-aware client the same call is issued by the LLM automatically once the server is configured — see integrations.

Result

Success responses are JSON of the shape below. Optional fields are omitted when the underlying data isn't present.

{
  "calls": [
    {
      "session_id": "a1b2c3d4",
      "agent_name": "Alpha Agent",
      "date": "2026-05-10T14:22:00Z",
      "duration_s": 412,
      "compass_score": 4.1,
      "lift": 0.22,
      "lift_band": "positive",
      "signal_count": 9,
      "summary": "Customer asked about plan upgrade; agent confirmed and scheduled a follow-up."
    }
  ],
  "count": 1
}

Auth context

Every call receives the connection's project_id and access_scope automatically — the caller never passes them. See Authentication & RBAC for how scope narrows results.

Errors

All tools return JSON. Errors are wrapped in { "error": "..." }. The most common shapes are documented in Python integration → Error shape.