Update project

Name of the project Example: "Default Project"

Outbound timeout in seconds Example: 300

Max call duration in seconds Example: 300

Min call duration in seconds Example: 30

Metric sampling rate (0.0 to 1.0). Default is 1.0 (100% of calls). Example: 0.5 (50%), 0.1 (10%), 0.001 (0.1%)

Max messages chat runs limit Example: 100

Webhook URL Example: "https://yourdomain.com/webhook"

Webhook secret Example: "your-webhook-secret"

Evaluate relevant metrics enabled Example: true or false

End call enabled Example: true or false

Generate scenario auto assign numbers enabled Example: true or false

Enable client side testing Example: true or false

Should show powered by Example: true or false

Notify on result completion Example: true or false

Notify daily report Example: true or false

Notify daily report even if there are no calls Example: true or false

Replace the default daily report data dump with an AI-narrated Top Issues section in Slack and email. Off by default. When on, Slack also exposes a "Show more" button that opens a modal with the legacy detail sections. Example: true or false

Notify cron failure Example: true or false

Notify cron success Example: true or false

Notify no calls Example: true or false

Configuration for no calls notification time periods, timeframes, exception numbers, and agents to exclude Example:

{
    "time_zone": "America/New_York",
    "time_periods": [
        {
            "start_time": "09:00",
            "end_time": "17:00",
            "time_frame": 30
        },
        {
            "start_time": "18:00",
            "end_time": "23:59",
            "time_frame": 60
        }
    ],
    "exception_numbers": ["+15551234567", "+14155552671"],
    "agents": [1, 10, 90]
}

Note:

• Calls with customer_number in exception_numbers list will be ignored when checking for no calls alert.
• If agents list is provided and not empty, those specific agents will be excluded from monitoring. All other agents in the project will be monitored. If omitted or empty, all agents in the project will be monitored.

Notify daily report webhook Example: true or false

Notify result webhook failed Example: true or false

Notify result webhook success Example: true or false

Notify result webhook for error runs (runs with error_message) Example: true or false

Notify cronjob webhook failed Example: true or false

Notify cronjob webhook success Example: true or false

Notify cronjob webhook for error runs (runs with error_message) Example: true or false

Notify call log webhook failed Example: true or false

Notify call log webhook success Example: true or false

Forward vapi webhook Example: true or false

Forward retell webhook Example: true or false

Is send emails enabled Example: true or false

Project-level scoring rubric for evaluating call success. Each call log produced under this project is run against the rubric and gets a single pass/fail outcome based on its metric results.

IMPORTANT (read before writing):

• NEVER blindly overwrite this field. Always first GET the project, read the existing rubric_config, then merge in only the rule(s) you intend to add/update, and PATCH the full merged value. Sending a partial value will silently delete every rule you did not include.

Structure:

• evaluation_logic: MUST be "and". All rules have to pass for a call to be marked successful. (Other values are not supported by the product currently.)
• rules: List of rules. Use one rule per (metric, condition) — keep each rule simple, with a single condition inside.

Each rule shape:

{
    "metric_id": <metric id>,
    "conditions": {
        "operator": "and",
        "conditions": [
            {"field": <field>, "op": <op>, "value": <value>}
        ]
    }
}

What field, op, and value to use depends on the metric type. The three supported cases:

1. Numeric / score metrics (qualitative scores, numeric metrics, workflow adherence with a score, etc.):

   • field: "score"
   • op: one of "gt", "gte", "lt", "lte", "eq"
   • value: a number (qualitative scores are on a 0–5 scale)
   • Example (score must be at least 4): {"field": "score", "op": "gte", "value": 4}

2. Boolean / binary metrics (eval_type binary_qualitative, binary_workflow_adherence, or evaluate_expected_outcome):

   • Boolean outcomes are stored as a score: True → 5, False → 0.
   • field: "score"
   • op: "eq" or "neq"
   • value: 5 for True, 0 for False
   • Example (must be True): {"field": "score", "op": "eq", "value": 5}
   • Example (must be False): {"field": "score", "op": "eq", "value": 0}

3. Enum metrics (eval_type enum, or observability_topics / observability_dropoff):

   • field: "enum"
   • op: "eq", "neq", or "in"
   • value: a single string for eq/neq, or a list of strings for in. Must match one of the metric's defined enum_values.
   • Example: {"field": "enum", "op": "eq", "value": "resolved"}
   • Example (any of): {"field": "enum", "op": "in", "value": ["resolved", "partially_resolved"]}

Full example:

{
    "evaluation_logic": "and",
    "rules": [
        {
            "metric_id": 115164,
            "conditions": {
                "operator": "and",
                "conditions": [
                    {"op": "eq", "field": "score", "value": 5}
                ]
            }
        },
        {
            "metric_id": 5512,
            "conditions": {
                "operator": "and",
                "conditions": [
                    {"op": "gte", "field": "score", "value": 4}
                ]
            }
        },
        {
            "metric_id": 8821,
            "conditions": {
                "operator": "and",
                "conditions": [
                    {"op": "in", "field": "enum", "value": ["resolved", "partially_resolved"]}
                ]
            }
        }
    ]
}

Notes:

• metric_id must reference a metric belonging to this project (project-scoped or agent-scoped under this project); unknown ids are rejected.
• Rules whose metric did not produce a value on a given call are skipped (neither pass nor fail) and do not affect the outcome.
• Pass an empty object {} to clear the rubric.

Project Timezone

VAPI API key Example: "vapi_api_key_123"

Retell API key Example: "retell_api_key_123"

Syntflow API key Example: "syntflow_api_key_123"

ElevenLabs API key Example: "elevenlabs_api_key_123"

LiveKit API key Example: "livekit_api_key_123"

Pipecat/Daily.co API key Example: "pipecat_api_key_123"

Name of the project Example: "Default Project"

Outbound timeout in seconds Example: 300

Max call duration in seconds Example: 300

Min call duration in seconds Example: 30

Metric sampling rate (0.0 to 1.0). Default is 1.0 (100% of calls). Example: 0.5 (50%), 0.1 (10%), 0.001 (0.1%)

Max messages chat runs limit Example: 100

Webhook URL Example: "https://yourdomain.com/webhook"

Webhook secret Example: "your-webhook-secret"

Evaluate relevant metrics enabled Example: true or false

End call enabled Example: true or false

Generate scenario auto assign numbers enabled Example: true or false

Enable client side testing Example: true or false

Should show powered by Example: true or false

Notify on result completion Example: true or false

Notify daily report Example: true or false

Notify daily report even if there are no calls Example: true or false

Replace the default daily report data dump with an AI-narrated Top Issues section in Slack and email. Off by default. When on, Slack also exposes a "Show more" button that opens a modal with the legacy detail sections. Example: true or false

Notify cron failure Example: true or false

Notify cron success Example: true or false

Notify no calls Example: true or false

Configuration for no calls notification time periods, timeframes, exception numbers, and agents to exclude Example:

{
    "time_zone": "America/New_York",
    "time_periods": [
        {
            "start_time": "09:00",
            "end_time": "17:00",
            "time_frame": 30
        },
        {
            "start_time": "18:00",
            "end_time": "23:59",
            "time_frame": 60
        }
    ],
    "exception_numbers": ["+15551234567", "+14155552671"],
    "agents": [1, 10, 90]
}

Note:

• Calls with customer_number in exception_numbers list will be ignored when checking for no calls alert.
• If agents list is provided and not empty, those specific agents will be excluded from monitoring. All other agents in the project will be monitored. If omitted or empty, all agents in the project will be monitored.

Notify daily report webhook Example: true or false

Notify result webhook failed Example: true or false

Notify result webhook success Example: true or false

Notify result webhook for error runs (runs with error_message) Example: true or false

Notify cronjob webhook failed Example: true or false

Notify cronjob webhook success Example: true or false

Notify cronjob webhook for error runs (runs with error_message) Example: true or false

Notify call log webhook failed Example: true or false

Notify call log webhook success Example: true or false

Forward vapi webhook Example: true or false

Forward retell webhook Example: true or false

Is send emails enabled Example: true or false

Project-level scoring rubric for evaluating call success. Each call log produced under this project is run against the rubric and gets a single pass/fail outcome based on its metric results.

IMPORTANT (read before writing):

• NEVER blindly overwrite this field. Always first GET the project, read the existing rubric_config, then merge in only the rule(s) you intend to add/update, and PATCH the full merged value. Sending a partial value will silently delete every rule you did not include.

Structure:

• evaluation_logic: MUST be "and". All rules have to pass for a call to be marked successful. (Other values are not supported by the product currently.)
• rules: List of rules. Use one rule per (metric, condition) — keep each rule simple, with a single condition inside.

Each rule shape:

{
    "metric_id": <metric id>,
    "conditions": {
        "operator": "and",
        "conditions": [
            {"field": <field>, "op": <op>, "value": <value>}
        ]
    }
}

What field, op, and value to use depends on the metric type. The three supported cases:

1. Numeric / score metrics (qualitative scores, numeric metrics, workflow adherence with a score, etc.):

   • field: "score"
   • op: one of "gt", "gte", "lt", "lte", "eq"
   • value: a number (qualitative scores are on a 0–5 scale)
   • Example (score must be at least 4): {"field": "score", "op": "gte", "value": 4}

2. Boolean / binary metrics (eval_type binary_qualitative, binary_workflow_adherence, or evaluate_expected_outcome):

   • Boolean outcomes are stored as a score: True → 5, False → 0.
   • field: "score"
   • op: "eq" or "neq"
   • value: 5 for True, 0 for False
   • Example (must be True): {"field": "score", "op": "eq", "value": 5}
   • Example (must be False): {"field": "score", "op": "eq", "value": 0}

3. Enum metrics (eval_type enum, or observability_topics / observability_dropoff):

   • field: "enum"
   • op: "eq", "neq", or "in"
   • value: a single string for eq/neq, or a list of strings for in. Must match one of the metric's defined enum_values.
   • Example: {"field": "enum", "op": "eq", "value": "resolved"}
   • Example (any of): {"field": "enum", "op": "in", "value": ["resolved", "partially_resolved"]}

Full example:

{
    "evaluation_logic": "and",
    "rules": [
        {
            "metric_id": 115164,
            "conditions": {
                "operator": "and",
                "conditions": [
                    {"op": "eq", "field": "score", "value": 5}
                ]
            }
        },
        {
            "metric_id": 5512,
            "conditions": {
                "operator": "and",
                "conditions": [
                    {"op": "gte", "field": "score", "value": 4}
                ]
            }
        },
        {
            "metric_id": 8821,
            "conditions": {
                "operator": "and",
                "conditions": [
                    {"op": "in", "field": "enum", "value": ["resolved", "partially_resolved"]}
                ]
            }
        }
    ]
}

Notes:

• metric_id must reference a metric belonging to this project (project-scoped or agent-scoped under this project); unknown ids are rejected.
• Rules whose metric did not produce a value on a given call are skipped (neither pass nor fail) and do not affect the outcome.
• Pass an empty object {} to clear the rubric.

Project Timezone