--- title: "Create Insight" method: POST path: "/reporting/insight" tags: ["Insight"] --- # Create Insight `POST /reporting/insight` ## Request body - union - CreateBarInsightFromCallTableDTO - `name` string — This is the name of the Insight. - `type` 'bar', required — This is the type of the Insight. It is required to be `bar` to create a bar insight. - `formulas` InsightFormula[] — Formulas are mathematical expressions applied on the data returned by the queries to transform them before being used to create the insight. The formulas needs to be a valid mathematical expression, supported by MathJS - https://mathjs.org/docs/expressions/syntax.html A formula is created by using the query names as the variable. The formulas must contain at least one query name in the LiquidJS format {{query_name}} or {{['query name']}} which will be substituted with the query result. For example, if you have 2 queries, 'Was Booking Made' and 'Average Call Duration', you can create a formula like this: ``` {{['Query 1']}} / {{['Query 2']}} * 100 ``` ``` ({{[Query 1]}} * 10) + {{[Query 2]}} ``` This will take the You can also use the query names as the variable in the formula. - `name` string — This is the name of the formula. It will be used to label the formula in the insight board on the UI. - `formula` string, required — This is the formula to calculate the insight from the queries. The formula needs to be a valid mathematical expression. The formula must contain at least one query name in the LiquidJS format {{query_name}} or {{['query name']}} which will be substituted with the query result. Any MathJS formula is allowed - https://mathjs.org/docs/expressions/syntax.html Common valid math operations are +, -, *, /, % - `metadata` BarInsightMetadata - `xAxisLabel` string - `yAxisLabel` string - `yAxisMin` number - `yAxisMax` number - `name` string - `timeRange` InsightTimeRangeWithStep - `step` 'minute' | 'hour' | 'day' | 'week' | 'month' | 'quarter' | 'year' — This is the group by step for aggregation. If not provided, defaults to group by day. - `start` object — This is the start date for the time range. Should be a valid ISO 8601 date-time string or relative time string. If not provided, defaults to the 7 days ago. Relative time strings of the format "-{number}{unit}" are allowed. Valid units are: - d: days - h: hours - w: weeks - m: months - y: years - `end` object — This is the end date for the time range. Should be a valid ISO 8601 date-time string or relative time string. If not provided, defaults to now. Relative time strings of the format "-{number}{unit}" are allowed. Valid units are: - d: days - h: hours - w: weeks - m: months - y: years - `timezone` string — This is the timezone you want to set for the query. If not provided, defaults to UTC. - `groupBy` 'assistantId' | 'workflowId' | 'squadId' | 'phoneNumberId' | 'type' | 'endedReason' | 'customerNumber' | 'campaignId' | 'artifact.structuredOutputs[OutputID]' — This is the group by column for the insight when table is `call`. These are the columns to group the results by. All results are grouped by the time range step by default. - `queries` union[], required — These are the queries to run to generate the insight. - union - JSONQueryOnCallTableWithStringTypeColumn - `type` 'vapiql-json', required — This is the type of query. Only allowed type is "vapiql-json". - `table` 'call', required — This is the table that will be queried. - `filters` union[] — This is the filters to apply to the insight. The discriminator automatically selects the correct filter type based on column and operator. - union - … - `column` 'id' | 'artifact.structuredOutputs[OutputID]', required — This is the column that will be queried in the selected table. Available columns depend on the selected table. String Type columns are columns where the rows store String data - `operation` 'count', required — This is the aggregation operation to perform on the column. When the column is a string type, the operation must be "count". - `name` string — This is the name of the query. It will be used to label the query in the insight board on the UI. - JSONQueryOnCallTableWithNumberTypeColumn - `type` 'vapiql-json', required — This is the type of query. Only allowed type is "vapiql-json". - `table` 'call', required — This is the table that will be queried. - `filters` union[] — This is the filters to apply to the insight. The discriminator automatically selects the correct filter type based on column and operator. - union - … - `column` 'cost' | 'duration' | 'averageModelLatency' | 'averageVoiceLatency' | 'averageTranscriberLatency' | 'averageTurnLatency' | 'averageEndpointingLatency' | 'artifact.structuredOutputs[OutputID]', required — This is the column that will be queried in the selected table. Available columns depend on the selected table. Number Type columns are columns where the rows store Number data - `operation` 'average' | 'sum' | 'min' | 'max', required — This is the aggregation operation to perform on the column. When the column is a number type, the operation must be one of the following: - average - sum - min - max - `name` string — This is the name of the query. It will be used to label the query in the insight board on the UI. - JSONQueryOnCallTableWithStructuredOutputColumn - `type` 'vapiql-json', required — This is the type of query. Only allowed type is "vapiql-json". - `table` 'call', required — This is the table that will be queried. - `filters` union[] — This is the filters to apply to the insight. The discriminator automatically selects the correct filter type based on column and operator. - union - … - `column` 'artifact.structuredOutputs[OutputID]', required — This is the column that will be queried in the call table. Structured Output Type columns are only to query on artifact.structuredOutputs[OutputID] column. - `operation` 'average' | 'count' | 'sum' | 'min' | 'max', required — This is the aggregation operation to perform on the column. When the column is a structured output type, the operation depends on the value of the structured output. If the structured output is a string or boolean, the operation must be "count". If the structured output is a number, the operation can be "average", "sum", "min", or "max". - `name` string — This is the name of the query. It will be used to label the query in the insight board on the UI. - JSONQueryOnEventsTable - `type` 'vapiql-json', required — This is the type of query. Only allowed type is "vapiql-json". - `table` 'events', required — This is the table that will be queried. Must be "events" for event-based insights. - `on` 'call.started' | 'call.ended' | 'call.inProgress' | 'call.queued' | 'call.transportConnected' | 'call.transportDisconnected' | 'call.transportReconnected' | 'call.transferInitiated' | 'call.transferCompleted' | 'call.transferFailed' | 'call.transferCancelled' | 'call.handoffInitiated' | 'call.handoffCompleted' | 'call.handoffFailed' | 'call.assistantSwapped' | 'call.assistantStarted' | 'call.customerJoined' | 'call.customerLeft' | 'call.controlReceived' | 'call.listenStarted' | 'call.recordingStarted' | 'call.recordingPaused' | 'call.recordingResumed' | 'call.voicemailDetected' | 'call.voicemailNotDetected' | 'call.dtmfReceived' | 'call.dtmfSent' | 'call.amdDetected' | 'call.hookTriggered' | 'call.hookSucceeded' | 'call.hookFailed' | 'call.statusReceived' | 'call.silenceTimeout' | 'call.microphoneTimeout' | 'call.maxDurationReached' | 'assistant.voice.requestStarted' | 'assistant.voice.requestSucceeded' | 'assistant.voice.requestFailed' | 'assistant.voice.connectionOpened' | 'assistant.voice.connectionClosed' | 'assistant.voice.firstAudioReceived' | 'assistant.voice.audioChunkReceived' | 'assistant.voice.generationSucceeded' | 'assistant.voice.generationFailed' | 'assistant.voice.textPushed' | 'assistant.voice.reconnecting' | 'assistant.voice.cleanup' | 'assistant.voice.clearing' | 'assistant.voice.voiceSwitched' | 'assistant.model.requestStarted' | 'assistant.model.requestSucceeded' | 'assistant.model.requestFailed' | 'assistant.model.requestAttemptStarted' | 'assistant.model.requestAttemptSucceeded' | 'assistant.model.requestAttemptFailed' | 'assistant.model.connectionOpened' | 'assistant.model.connectionClosed' | 'assistant.model.firstTokenReceived' | 'assistant.model.tokenReceived' | 'assistant.model.responseSucceeded' | 'assistant.model.responseFailed' | 'assistant.model.toolCallsReceived' | 'assistant.model.reconnecting' | 'assistant.model.cleanup' | 'assistant.model.clearing' | 'assistant.tool.started' | 'assistant.tool.completed' | 'assistant.tool.failed' | 'assistant.tool.delayedMessageSent' | 'assistant.tool.timeout' | 'assistant.tool.asyncCallbackReceived' | 'assistant.transcriber.requestStarted' | 'assistant.transcriber.requestSucceeded' | 'assistant.transcriber.requestFailed' | 'assistant.transcriber.connectionOpened' | 'assistant.transcriber.connectionClosed' | 'assistant.transcriber.partialTranscript' | 'assistant.transcriber.finalTranscript' | 'assistant.transcriber.keepAlive' | 'assistant.transcriber.reconnecting' | 'assistant.transcriber.cleanup' | 'assistant.transcriber.clearing' | 'assistant.transcriber.transcriptIgnored' | 'assistant.transcriber.languageSwitched' | 'assistant.analysis.structuredOutputGenerated' | 'assistant.analysis.structuredOutputSkipped' | 'pipeline.turnStarted' | 'pipeline.cleared' | 'pipeline.botSpeechStarted' | 'pipeline.botSpeechStopped' | 'pipeline.userSpeechStarted' | 'pipeline.userSpeechStopped' | 'pipeline.endpointingTriggered' | 'pipeline.firstMessageStarted' | 'pipeline.firstMessageCompleted', required — The event type to query - `operation` 'count' | 'percentage', required — This is the operation to perform on matching events. - "count": Returns the raw count of matching events - "percentage": Returns (count of matching events / total calls) * 100 - `filters` union[] — These are the filters to apply to the events query. Each filter filters on a field specific to the event type. - union - … - `name` string — This is the name of the query. It will be used to label the query in the insight board on the UI. - CreatePieInsightFromCallTableDTO - `name` string — This is the name of the Insight. - `type` 'pie', required — This is the type of the Insight. It is required to be `pie` to create a pie insight. - `formulas` InsightFormula[] — Formulas are mathematical expressions applied on the data returned by the queries to transform them before being used to create the insight. The formulas needs to be a valid mathematical expression, supported by MathJS - https://mathjs.org/docs/expressions/syntax.html A formula is created by using the query names as the variable. The formulas must contain at least one query name in the LiquidJS format {{query_name}} or {{['query name']}} which will be substituted with the query result. For example, if you have 2 queries, 'Was Booking Made' and 'Average Call Duration', you can create a formula like this: ``` {{['Query 1']}} / {{['Query 2']}} * 100 ``` ``` ({{[Query 1]}} * 10) + {{[Query 2]}} ``` This will take the You can also use the query names as the variable in the formula. - `name` string — This is the name of the formula. It will be used to label the formula in the insight board on the UI. - `formula` string, required — This is the formula to calculate the insight from the queries. The formula needs to be a valid mathematical expression. The formula must contain at least one query name in the LiquidJS format {{query_name}} or {{['query name']}} which will be substituted with the query result. Any MathJS formula is allowed - https://mathjs.org/docs/expressions/syntax.html Common valid math operations are +, -, *, /, % - `timeRange` InsightTimeRange - `start` object — This is the start date for the time range. Should be a valid ISO 8601 date-time string or relative time string. If not provided, defaults to the 7 days ago. Relative time strings of the format "-{number}{unit}" are allowed. Valid units are: - d: days - h: hours - w: weeks - m: months - y: years - `end` object — This is the end date for the time range. Should be a valid ISO 8601 date-time string or relative time string. If not provided, defaults to now. Relative time strings of the format "-{number}{unit}" are allowed. Valid units are: - d: days - h: hours - w: weeks - m: months - y: years - `timezone` string — This is the timezone you want to set for the query. If not provided, defaults to UTC. - `groupBy` 'assistantId' | 'workflowId' | 'squadId' | 'phoneNumberId' | 'type' | 'endedReason' | 'customerNumber' | 'campaignId' | 'artifact.structuredOutputs[OutputID]' — This is the group by column for the insight when table is `call`. These are the columns to group the results by. All results are grouped by the time range step by default. - `queries` union[], required — These are the queries to run to generate the insight. - union - JSONQueryOnCallTableWithStringTypeColumn - `type` 'vapiql-json', required — This is the type of query. Only allowed type is "vapiql-json". - `table` 'call', required — This is the table that will be queried. - `filters` union[] — This is the filters to apply to the insight. The discriminator automatically selects the correct filter type based on column and operator. - union - … - `column` 'id' | 'artifact.structuredOutputs[OutputID]', required — This is the column that will be queried in the selected table. Available columns depend on the selected table. String Type columns are columns where the rows store String data - `operation` 'count', required — This is the aggregation operation to perform on the column. When the column is a string type, the operation must be "count". - `name` string — This is the name of the query. It will be used to label the query in the insight board on the UI. - JSONQueryOnCallTableWithNumberTypeColumn - `type` 'vapiql-json', required — This is the type of query. Only allowed type is "vapiql-json". - `table` 'call', required — This is the table that will be queried. - `filters` union[] — This is the filters to apply to the insight. The discriminator automatically selects the correct filter type based on column and operator. - union - … - `column` 'cost' | 'duration' | 'averageModelLatency' | 'averageVoiceLatency' | 'averageTranscriberLatency' | 'averageTurnLatency' | 'averageEndpointingLatency' | 'artifact.structuredOutputs[OutputID]', required — This is the column that will be queried in the selected table. Available columns depend on the selected table. Number Type columns are columns where the rows store Number data - `operation` 'average' | 'sum' | 'min' | 'max', required — This is the aggregation operation to perform on the column. When the column is a number type, the operation must be one of the following: - average - sum - min - max - `name` string — This is the name of the query. It will be used to label the query in the insight board on the UI. - JSONQueryOnCallTableWithStructuredOutputColumn - `type` 'vapiql-json', required — This is the type of query. Only allowed type is "vapiql-json". - `table` 'call', required — This is the table that will be queried. - `filters` union[] — This is the filters to apply to the insight. The discriminator automatically selects the correct filter type based on column and operator. - union - … - `column` 'artifact.structuredOutputs[OutputID]', required — This is the column that will be queried in the call table. Structured Output Type columns are only to query on artifact.structuredOutputs[OutputID] column. - `operation` 'average' | 'count' | 'sum' | 'min' | 'max', required — This is the aggregation operation to perform on the column. When the column is a structured output type, the operation depends on the value of the structured output. If the structured output is a string or boolean, the operation must be "count". If the structured output is a number, the operation can be "average", "sum", "min", or "max". - `name` string — This is the name of the query. It will be used to label the query in the insight board on the UI. - CreateLineInsightFromCallTableDTO - `name` string — This is the name of the Insight. - `type` 'line', required — This is the type of the Insight. It is required to be `line` to create a line insight. - `formulas` InsightFormula[] — Formulas are mathematical expressions applied on the data returned by the queries to transform them before being used to create the insight. The formulas needs to be a valid mathematical expression, supported by MathJS - https://mathjs.org/docs/expressions/syntax.html A formula is created by using the query names as the variable. The formulas must contain at least one query name in the LiquidJS format {{query_name}} or {{['query name']}} which will be substituted with the query result. For example, if you have 2 queries, 'Was Booking Made' and 'Average Call Duration', you can create a formula like this: ``` {{['Query 1']}} / {{['Query 2']}} * 100 ``` ``` ({{[Query 1]}} * 10) + {{[Query 2]}} ``` This will take the You can also use the query names as the variable in the formula. - `name` string — This is the name of the formula. It will be used to label the formula in the insight board on the UI. - `formula` string, required — This is the formula to calculate the insight from the queries. The formula needs to be a valid mathematical expression. The formula must contain at least one query name in the LiquidJS format {{query_name}} or {{['query name']}} which will be substituted with the query result. Any MathJS formula is allowed - https://mathjs.org/docs/expressions/syntax.html Common valid math operations are +, -, *, /, % - `metadata` LineInsightMetadata - `xAxisLabel` string - `yAxisLabel` string - `yAxisMin` number - `yAxisMax` number - `name` string - `timeRange` InsightTimeRangeWithStep - `step` 'minute' | 'hour' | 'day' | 'week' | 'month' | 'quarter' | 'year' — This is the group by step for aggregation. If not provided, defaults to group by day. - `start` object — This is the start date for the time range. Should be a valid ISO 8601 date-time string or relative time string. If not provided, defaults to the 7 days ago. Relative time strings of the format "-{number}{unit}" are allowed. Valid units are: - d: days - h: hours - w: weeks - m: months - y: years - `end` object — This is the end date for the time range. Should be a valid ISO 8601 date-time string or relative time string. If not provided, defaults to now. Relative time strings of the format "-{number}{unit}" are allowed. Valid units are: - d: days - h: hours - w: weeks - m: months - y: years - `timezone` string — This is the timezone you want to set for the query. If not provided, defaults to UTC. - `groupBy` 'assistantId' | 'workflowId' | 'squadId' | 'phoneNumberId' | 'type' | 'endedReason' | 'customerNumber' | 'campaignId' | 'artifact.structuredOutputs[OutputID]' — This is the group by column for the insight when table is `call`. These are the columns to group the results by. All results are grouped by the time range step by default. - `queries` union[], required — These are the queries to run to generate the insight. - union - JSONQueryOnCallTableWithStringTypeColumn - `type` 'vapiql-json', required — This is the type of query. Only allowed type is "vapiql-json". - `table` 'call', required — This is the table that will be queried. - `filters` union[] — This is the filters to apply to the insight. The discriminator automatically selects the correct filter type based on column and operator. - union - … - `column` 'id' | 'artifact.structuredOutputs[OutputID]', required — This is the column that will be queried in the selected table. Available columns depend on the selected table. String Type columns are columns where the rows store String data - `operation` 'count', required — This is the aggregation operation to perform on the column. When the column is a string type, the operation must be "count". - `name` string — This is the name of the query. It will be used to label the query in the insight board on the UI. - JSONQueryOnCallTableWithNumberTypeColumn - `type` 'vapiql-json', required — This is the type of query. Only allowed type is "vapiql-json". - `table` 'call', required — This is the table that will be queried. - `filters` union[] — This is the filters to apply to the insight. The discriminator automatically selects the correct filter type based on column and operator. - union - … - `column` 'cost' | 'duration' | 'averageModelLatency' | 'averageVoiceLatency' | 'averageTranscriberLatency' | 'averageTurnLatency' | 'averageEndpointingLatency' | 'artifact.structuredOutputs[OutputID]', required — This is the column that will be queried in the selected table. Available columns depend on the selected table. Number Type columns are columns where the rows store Number data - `operation` 'average' | 'sum' | 'min' | 'max', required — This is the aggregation operation to perform on the column. When the column is a number type, the operation must be one of the following: - average - sum - min - max - `name` string — This is the name of the query. It will be used to label the query in the insight board on the UI. - JSONQueryOnCallTableWithStructuredOutputColumn - `type` 'vapiql-json', required — This is the type of query. Only allowed type is "vapiql-json". - `table` 'call', required — This is the table that will be queried. - `filters` union[] — This is the filters to apply to the insight. The discriminator automatically selects the correct filter type based on column and operator. - union - … - `column` 'artifact.structuredOutputs[OutputID]', required — This is the column that will be queried in the call table. Structured Output Type columns are only to query on artifact.structuredOutputs[OutputID] column. - `operation` 'average' | 'count' | 'sum' | 'min' | 'max', required — This is the aggregation operation to perform on the column. When the column is a structured output type, the operation depends on the value of the structured output. If the structured output is a string or boolean, the operation must be "count". If the structured output is a number, the operation can be "average", "sum", "min", or "max". - `name` string — This is the name of the query. It will be used to label the query in the insight board on the UI. - CreateTextInsightFromCallTableDTO - `name` string — This is the name of the Insight. - `type` 'text', required — This is the type of the Insight. It is required to be `text` to create a text insight. - `formula` object — Formulas are mathematical expressions applied on the data returned by the queries to transform them before being used to create the insight. The formulas needs to be a valid mathematical expression, supported by MathJS - https://mathjs.org/docs/expressions/syntax.html A formula is created by using the query names as the variable. The formulas must contain at least one query name in the LiquidJS format {{query_name}} or {{['query name']}} which will be substituted with the query result. For example, if you have 2 queries, 'Was Booking Made' and 'Average Call Duration', you can create a formula like this: ``` {{['Query 1']}} / {{['Query 2']}} * 100 ``` ``` ({{[Query 1]}} * 10) + {{[Query 2]}} ``` This will take the You can also use the query names as the variable in the formula. - `timeRange` InsightTimeRange - `start` object — This is the start date for the time range. Should be a valid ISO 8601 date-time string or relative time string. If not provided, defaults to the 7 days ago. Relative time strings of the format "-{number}{unit}" are allowed. Valid units are: - d: days - h: hours - w: weeks - m: months - y: years - `end` object — This is the end date for the time range. Should be a valid ISO 8601 date-time string or relative time string. If not provided, defaults to now. Relative time strings of the format "-{number}{unit}" are allowed. Valid units are: - d: days - h: hours - w: weeks - m: months - y: years - `timezone` string — This is the timezone you want to set for the query. If not provided, defaults to UTC. - `queries` union[], required — These are the queries to run to generate the insight. For Text Insights, we only allow a single query, or require a formula if multiple queries are provided - union - JSONQueryOnCallTableWithStringTypeColumn - `type` 'vapiql-json', required — This is the type of query. Only allowed type is "vapiql-json". - `table` 'call', required — This is the table that will be queried. - `filters` union[] — This is the filters to apply to the insight. The discriminator automatically selects the correct filter type based on column and operator. - union - … - `column` 'id' | 'artifact.structuredOutputs[OutputID]', required — This is the column that will be queried in the selected table. Available columns depend on the selected table. String Type columns are columns where the rows store String data - `operation` 'count', required — This is the aggregation operation to perform on the column. When the column is a string type, the operation must be "count". - `name` string — This is the name of the query. It will be used to label the query in the insight board on the UI. - JSONQueryOnCallTableWithNumberTypeColumn - `type` 'vapiql-json', required — This is the type of query. Only allowed type is "vapiql-json". - `table` 'call', required — This is the table that will be queried. - `filters` union[] — This is the filters to apply to the insight. The discriminator automatically selects the correct filter type based on column and operator. - union - … - `column` 'cost' | 'duration' | 'averageModelLatency' | 'averageVoiceLatency' | 'averageTranscriberLatency' | 'averageTurnLatency' | 'averageEndpointingLatency' | 'artifact.structuredOutputs[OutputID]', required — This is the column that will be queried in the selected table. Available columns depend on the selected table. Number Type columns are columns where the rows store Number data - `operation` 'average' | 'sum' | 'min' | 'max', required — This is the aggregation operation to perform on the column. When the column is a number type, the operation must be one of the following: - average - sum - min - max - `name` string — This is the name of the query. It will be used to label the query in the insight board on the UI. - JSONQueryOnCallTableWithStructuredOutputColumn - `type` 'vapiql-json', required — This is the type of query. Only allowed type is "vapiql-json". - `table` 'call', required — This is the table that will be queried. - `filters` union[] — This is the filters to apply to the insight. The discriminator automatically selects the correct filter type based on column and operator. - union - … - `column` 'artifact.structuredOutputs[OutputID]', required — This is the column that will be queried in the call table. Structured Output Type columns are only to query on artifact.structuredOutputs[OutputID] column. - `operation` 'average' | 'count' | 'sum' | 'min' | 'max', required — This is the aggregation operation to perform on the column. When the column is a structured output type, the operation depends on the value of the structured output. If the structured output is a string or boolean, the operation must be "count". If the structured output is a number, the operation can be "average", "sum", "min", or "max". - `name` string — This is the name of the query. It will be used to label the query in the insight board on the UI. ## Response `201` - union - BarInsight - `name` string — This is the name of the Insight. - `type` 'bar', required — This is the type of the Insight. It is required to be `bar` to create a bar insight. - `formulas` InsightFormula[] — Formulas are mathematical expressions applied on the data returned by the queries to transform them before being used to create the insight. The formulas needs to be a valid mathematical expression, supported by MathJS - https://mathjs.org/docs/expressions/syntax.html A formula is created by using the query names as the variable. The formulas must contain at least one query name in the LiquidJS format {{query_name}} or {{['query name']}} which will be substituted with the query result. For example, if you have 2 queries, 'Was Booking Made' and 'Average Call Duration', you can create a formula like this: ``` {{['Query 1']}} / {{['Query 2']}} * 100 ``` ``` ({{[Query 1]}} * 10) + {{[Query 2]}} ``` This will take the You can also use the query names as the variable in the formula. - `name` string — This is the name of the formula. It will be used to label the formula in the insight board on the UI. - `formula` string, required — This is the formula to calculate the insight from the queries. The formula needs to be a valid mathematical expression. The formula must contain at least one query name in the LiquidJS format {{query_name}} or {{['query name']}} which will be substituted with the query result. Any MathJS formula is allowed - https://mathjs.org/docs/expressions/syntax.html Common valid math operations are +, -, *, /, % - `metadata` BarInsightMetadata - `xAxisLabel` string - `yAxisLabel` string - `yAxisMin` number - `yAxisMax` number - `name` string - `timeRange` InsightTimeRangeWithStep - `step` 'minute' | 'hour' | 'day' | 'week' | 'month' | 'quarter' | 'year' — This is the group by step for aggregation. If not provided, defaults to group by day. - `start` object — This is the start date for the time range. Should be a valid ISO 8601 date-time string or relative time string. If not provided, defaults to the 7 days ago. Relative time strings of the format "-{number}{unit}" are allowed. Valid units are: - d: days - h: hours - w: weeks - m: months - y: years - `end` object — This is the end date for the time range. Should be a valid ISO 8601 date-time string or relative time string. If not provided, defaults to now. Relative time strings of the format "-{number}{unit}" are allowed. Valid units are: - d: days - h: hours - w: weeks - m: months - y: years - `timezone` string — This is the timezone you want to set for the query. If not provided, defaults to UTC. - `groupBy` 'assistantId' | 'workflowId' | 'squadId' | 'phoneNumberId' | 'type' | 'endedReason' | 'customerNumber' | 'campaignId' | 'artifact.structuredOutputs[OutputID]' — This is the group by column for the insight when table is `call`. These are the columns to group the results by. All results are grouped by the time range step by default. - `queries` union[], required — These are the queries to run to generate the insight. - union - JSONQueryOnCallTableWithStringTypeColumn - `type` 'vapiql-json', required — This is the type of query. Only allowed type is "vapiql-json". - `table` 'call', required — This is the table that will be queried. - `filters` union[] — This is the filters to apply to the insight. The discriminator automatically selects the correct filter type based on column and operator. - union - … - `column` 'id' | 'artifact.structuredOutputs[OutputID]', required — This is the column that will be queried in the selected table. Available columns depend on the selected table. String Type columns are columns where the rows store String data - `operation` 'count', required — This is the aggregation operation to perform on the column. When the column is a string type, the operation must be "count". - `name` string — This is the name of the query. It will be used to label the query in the insight board on the UI. - JSONQueryOnCallTableWithNumberTypeColumn - `type` 'vapiql-json', required — This is the type of query. Only allowed type is "vapiql-json". - `table` 'call', required — This is the table that will be queried. - `filters` union[] — This is the filters to apply to the insight. The discriminator automatically selects the correct filter type based on column and operator. - union - … - `column` 'cost' | 'duration' | 'averageModelLatency' | 'averageVoiceLatency' | 'averageTranscriberLatency' | 'averageTurnLatency' | 'averageEndpointingLatency' | 'artifact.structuredOutputs[OutputID]', required — This is the column that will be queried in the selected table. Available columns depend on the selected table. Number Type columns are columns where the rows store Number data - `operation` 'average' | 'sum' | 'min' | 'max', required — This is the aggregation operation to perform on the column. When the column is a number type, the operation must be one of the following: - average - sum - min - max - `name` string — This is the name of the query. It will be used to label the query in the insight board on the UI. - JSONQueryOnCallTableWithStructuredOutputColumn - `type` 'vapiql-json', required — This is the type of query. Only allowed type is "vapiql-json". - `table` 'call', required — This is the table that will be queried. - `filters` union[] — This is the filters to apply to the insight. The discriminator automatically selects the correct filter type based on column and operator. - union - … - `column` 'artifact.structuredOutputs[OutputID]', required — This is the column that will be queried in the call table. Structured Output Type columns are only to query on artifact.structuredOutputs[OutputID] column. - `operation` 'average' | 'count' | 'sum' | 'min' | 'max', required — This is the aggregation operation to perform on the column. When the column is a structured output type, the operation depends on the value of the structured output. If the structured output is a string or boolean, the operation must be "count". If the structured output is a number, the operation can be "average", "sum", "min", or "max". - `name` string — This is the name of the query. It will be used to label the query in the insight board on the UI. - JSONQueryOnEventsTable - `type` 'vapiql-json', required — This is the type of query. Only allowed type is "vapiql-json". - `table` 'events', required — This is the table that will be queried. Must be "events" for event-based insights. - `on` 'call.started' | 'call.ended' | 'call.inProgress' | 'call.queued' | 'call.transportConnected' | 'call.transportDisconnected' | 'call.transportReconnected' | 'call.transferInitiated' | 'call.transferCompleted' | 'call.transferFailed' | 'call.transferCancelled' | 'call.handoffInitiated' | 'call.handoffCompleted' | 'call.handoffFailed' | 'call.assistantSwapped' | 'call.assistantStarted' | 'call.customerJoined' | 'call.customerLeft' | 'call.controlReceived' | 'call.listenStarted' | 'call.recordingStarted' | 'call.recordingPaused' | 'call.recordingResumed' | 'call.voicemailDetected' | 'call.voicemailNotDetected' | 'call.dtmfReceived' | 'call.dtmfSent' | 'call.amdDetected' | 'call.hookTriggered' | 'call.hookSucceeded' | 'call.hookFailed' | 'call.statusReceived' | 'call.silenceTimeout' | 'call.microphoneTimeout' | 'call.maxDurationReached' | 'assistant.voice.requestStarted' | 'assistant.voice.requestSucceeded' | 'assistant.voice.requestFailed' | 'assistant.voice.connectionOpened' | 'assistant.voice.connectionClosed' | 'assistant.voice.firstAudioReceived' | 'assistant.voice.audioChunkReceived' | 'assistant.voice.generationSucceeded' | 'assistant.voice.generationFailed' | 'assistant.voice.textPushed' | 'assistant.voice.reconnecting' | 'assistant.voice.cleanup' | 'assistant.voice.clearing' | 'assistant.voice.voiceSwitched' | 'assistant.model.requestStarted' | 'assistant.model.requestSucceeded' | 'assistant.model.requestFailed' | 'assistant.model.requestAttemptStarted' | 'assistant.model.requestAttemptSucceeded' | 'assistant.model.requestAttemptFailed' | 'assistant.model.connectionOpened' | 'assistant.model.connectionClosed' | 'assistant.model.firstTokenReceived' | 'assistant.model.tokenReceived' | 'assistant.model.responseSucceeded' | 'assistant.model.responseFailed' | 'assistant.model.toolCallsReceived' | 'assistant.model.reconnecting' | 'assistant.model.cleanup' | 'assistant.model.clearing' | 'assistant.tool.started' | 'assistant.tool.completed' | 'assistant.tool.failed' | 'assistant.tool.delayedMessageSent' | 'assistant.tool.timeout' | 'assistant.tool.asyncCallbackReceived' | 'assistant.transcriber.requestStarted' | 'assistant.transcriber.requestSucceeded' | 'assistant.transcriber.requestFailed' | 'assistant.transcriber.connectionOpened' | 'assistant.transcriber.connectionClosed' | 'assistant.transcriber.partialTranscript' | 'assistant.transcriber.finalTranscript' | 'assistant.transcriber.keepAlive' | 'assistant.transcriber.reconnecting' | 'assistant.transcriber.cleanup' | 'assistant.transcriber.clearing' | 'assistant.transcriber.transcriptIgnored' | 'assistant.transcriber.languageSwitched' | 'assistant.analysis.structuredOutputGenerated' | 'assistant.analysis.structuredOutputSkipped' | 'pipeline.turnStarted' | 'pipeline.cleared' | 'pipeline.botSpeechStarted' | 'pipeline.botSpeechStopped' | 'pipeline.userSpeechStarted' | 'pipeline.userSpeechStopped' | 'pipeline.endpointingTriggered' | 'pipeline.firstMessageStarted' | 'pipeline.firstMessageCompleted', required — The event type to query - `operation` 'count' | 'percentage', required — This is the operation to perform on matching events. - "count": Returns the raw count of matching events - "percentage": Returns (count of matching events / total calls) * 100 - `filters` union[] — These are the filters to apply to the events query. Each filter filters on a field specific to the event type. - union - … - `name` string — This is the name of the query. It will be used to label the query in the insight board on the UI. - `id` string, required — This is the unique identifier for the Insight. - `orgId` string, required — This is the unique identifier for the org that this Insight belongs to. - `createdAt` string, date-time, required — This is the ISO 8601 date-time string of when the Insight was created. - `updatedAt` string, date-time, required — This is the ISO 8601 date-time string of when the Insight was last updated. - `systemKey` string — Stable server-owned identifier for system-created insights. - PieInsight - `name` string — This is the name of the Insight. - `type` 'pie', required — This is the type of the Insight. It is required to be `pie` to create a pie insight. - `formulas` InsightFormula[] — Formulas are mathematical expressions applied on the data returned by the queries to transform them before being used to create the insight. The formulas needs to be a valid mathematical expression, supported by MathJS - https://mathjs.org/docs/expressions/syntax.html A formula is created by using the query names as the variable. The formulas must contain at least one query name in the LiquidJS format {{query_name}} or {{['query name']}} which will be substituted with the query result. For example, if you have 2 queries, 'Was Booking Made' and 'Average Call Duration', you can create a formula like this: ``` {{['Query 1']}} / {{['Query 2']}} * 100 ``` ``` ({{[Query 1]}} * 10) + {{[Query 2]}} ``` This will take the You can also use the query names as the variable in the formula. - `name` string — This is the name of the formula. It will be used to label the formula in the insight board on the UI. - `formula` string, required — This is the formula to calculate the insight from the queries. The formula needs to be a valid mathematical expression. The formula must contain at least one query name in the LiquidJS format {{query_name}} or {{['query name']}} which will be substituted with the query result. Any MathJS formula is allowed - https://mathjs.org/docs/expressions/syntax.html Common valid math operations are +, -, *, /, % - `timeRange` InsightTimeRange - `start` object — This is the start date for the time range. Should be a valid ISO 8601 date-time string or relative time string. If not provided, defaults to the 7 days ago. Relative time strings of the format "-{number}{unit}" are allowed. Valid units are: - d: days - h: hours - w: weeks - m: months - y: years - `end` object — This is the end date for the time range. Should be a valid ISO 8601 date-time string or relative time string. If not provided, defaults to now. Relative time strings of the format "-{number}{unit}" are allowed. Valid units are: - d: days - h: hours - w: weeks - m: months - y: years - `timezone` string — This is the timezone you want to set for the query. If not provided, defaults to UTC. - `groupBy` 'assistantId' | 'workflowId' | 'squadId' | 'phoneNumberId' | 'type' | 'endedReason' | 'customerNumber' | 'campaignId' | 'artifact.structuredOutputs[OutputID]' — This is the group by column for the insight when table is `call`. These are the columns to group the results by. All results are grouped by the time range step by default. - `queries` union[], required — These are the queries to run to generate the insight. - union - JSONQueryOnCallTableWithStringTypeColumn - `type` 'vapiql-json', required — This is the type of query. Only allowed type is "vapiql-json". - `table` 'call', required — This is the table that will be queried. - `filters` union[] — This is the filters to apply to the insight. The discriminator automatically selects the correct filter type based on column and operator. - union - … - `column` 'id' | 'artifact.structuredOutputs[OutputID]', required — This is the column that will be queried in the selected table. Available columns depend on the selected table. String Type columns are columns where the rows store String data - `operation` 'count', required — This is the aggregation operation to perform on the column. When the column is a string type, the operation must be "count". - `name` string — This is the name of the query. It will be used to label the query in the insight board on the UI. - JSONQueryOnCallTableWithNumberTypeColumn - `type` 'vapiql-json', required — This is the type of query. Only allowed type is "vapiql-json". - `table` 'call', required — This is the table that will be queried. - `filters` union[] — This is the filters to apply to the insight. The discriminator automatically selects the correct filter type based on column and operator. - union - … - `column` 'cost' | 'duration' | 'averageModelLatency' | 'averageVoiceLatency' | 'averageTranscriberLatency' | 'averageTurnLatency' | 'averageEndpointingLatency' | 'artifact.structuredOutputs[OutputID]', required — This is the column that will be queried in the selected table. Available columns depend on the selected table. Number Type columns are columns where the rows store Number data - `operation` 'average' | 'sum' | 'min' | 'max', required — This is the aggregation operation to perform on the column. When the column is a number type, the operation must be one of the following: - average - sum - min - max - `name` string — This is the name of the query. It will be used to label the query in the insight board on the UI. - JSONQueryOnCallTableWithStructuredOutputColumn - `type` 'vapiql-json', required — This is the type of query. Only allowed type is "vapiql-json". - `table` 'call', required — This is the table that will be queried. - `filters` union[] — This is the filters to apply to the insight. The discriminator automatically selects the correct filter type based on column and operator. - union - … - `column` 'artifact.structuredOutputs[OutputID]', required — This is the column that will be queried in the call table. Structured Output Type columns are only to query on artifact.structuredOutputs[OutputID] column. - `operation` 'average' | 'count' | 'sum' | 'min' | 'max', required — This is the aggregation operation to perform on the column. When the column is a structured output type, the operation depends on the value of the structured output. If the structured output is a string or boolean, the operation must be "count". If the structured output is a number, the operation can be "average", "sum", "min", or "max". - `name` string — This is the name of the query. It will be used to label the query in the insight board on the UI. - `id` string, required — This is the unique identifier for the Insight. - `orgId` string, required — This is the unique identifier for the org that this Insight belongs to. - `createdAt` string, date-time, required — This is the ISO 8601 date-time string of when the Insight was created. - `updatedAt` string, date-time, required — This is the ISO 8601 date-time string of when the Insight was last updated. - `systemKey` string — Stable server-owned identifier for system-created insights. - LineInsight - `name` string — This is the name of the Insight. - `type` 'line', required — This is the type of the Insight. It is required to be `line` to create a line insight. - `formulas` InsightFormula[] — Formulas are mathematical expressions applied on the data returned by the queries to transform them before being used to create the insight. The formulas needs to be a valid mathematical expression, supported by MathJS - https://mathjs.org/docs/expressions/syntax.html A formula is created by using the query names as the variable. The formulas must contain at least one query name in the LiquidJS format {{query_name}} or {{['query name']}} which will be substituted with the query result. For example, if you have 2 queries, 'Was Booking Made' and 'Average Call Duration', you can create a formula like this: ``` {{['Query 1']}} / {{['Query 2']}} * 100 ``` ``` ({{[Query 1]}} * 10) + {{[Query 2]}} ``` This will take the You can also use the query names as the variable in the formula. - `name` string — This is the name of the formula. It will be used to label the formula in the insight board on the UI. - `formula` string, required — This is the formula to calculate the insight from the queries. The formula needs to be a valid mathematical expression. The formula must contain at least one query name in the LiquidJS format {{query_name}} or {{['query name']}} which will be substituted with the query result. Any MathJS formula is allowed - https://mathjs.org/docs/expressions/syntax.html Common valid math operations are +, -, *, /, % - `metadata` LineInsightMetadata - `xAxisLabel` string - `yAxisLabel` string - `yAxisMin` number - `yAxisMax` number - `name` string - `timeRange` InsightTimeRangeWithStep - `step` 'minute' | 'hour' | 'day' | 'week' | 'month' | 'quarter' | 'year' — This is the group by step for aggregation. If not provided, defaults to group by day. - `start` object — This is the start date for the time range. Should be a valid ISO 8601 date-time string or relative time string. If not provided, defaults to the 7 days ago. Relative time strings of the format "-{number}{unit}" are allowed. Valid units are: - d: days - h: hours - w: weeks - m: months - y: years - `end` object — This is the end date for the time range. Should be a valid ISO 8601 date-time string or relative time string. If not provided, defaults to now. Relative time strings of the format "-{number}{unit}" are allowed. Valid units are: - d: days - h: hours - w: weeks - m: months - y: years - `timezone` string — This is the timezone you want to set for the query. If not provided, defaults to UTC. - `groupBy` 'assistantId' | 'workflowId' | 'squadId' | 'phoneNumberId' | 'type' | 'endedReason' | 'customerNumber' | 'campaignId' | 'artifact.structuredOutputs[OutputID]' — This is the group by column for the insight when table is `call`. These are the columns to group the results by. All results are grouped by the time range step by default. - `queries` union[], required — These are the queries to run to generate the insight. - union - JSONQueryOnCallTableWithStringTypeColumn - `type` 'vapiql-json', required — This is the type of query. Only allowed type is "vapiql-json". - `table` 'call', required — This is the table that will be queried. - `filters` union[] — This is the filters to apply to the insight. The discriminator automatically selects the correct filter type based on column and operator. - union - … - `column` 'id' | 'artifact.structuredOutputs[OutputID]', required — This is the column that will be queried in the selected table. Available columns depend on the selected table. String Type columns are columns where the rows store String data - `operation` 'count', required — This is the aggregation operation to perform on the column. When the column is a string type, the operation must be "count". - `name` string — This is the name of the query. It will be used to label the query in the insight board on the UI. - JSONQueryOnCallTableWithNumberTypeColumn - `type` 'vapiql-json', required — This is the type of query. Only allowed type is "vapiql-json". - `table` 'call', required — This is the table that will be queried. - `filters` union[] — This is the filters to apply to the insight. The discriminator automatically selects the correct filter type based on column and operator. - union - … - `column` 'cost' | 'duration' | 'averageModelLatency' | 'averageVoiceLatency' | 'averageTranscriberLatency' | 'averageTurnLatency' | 'averageEndpointingLatency' | 'artifact.structuredOutputs[OutputID]', required — This is the column that will be queried in the selected table. Available columns depend on the selected table. Number Type columns are columns where the rows store Number data - `operation` 'average' | 'sum' | 'min' | 'max', required — This is the aggregation operation to perform on the column. When the column is a number type, the operation must be one of the following: - average - sum - min - max - `name` string — This is the name of the query. It will be used to label the query in the insight board on the UI. - JSONQueryOnCallTableWithStructuredOutputColumn - `type` 'vapiql-json', required — This is the type of query. Only allowed type is "vapiql-json". - `table` 'call', required — This is the table that will be queried. - `filters` union[] — This is the filters to apply to the insight. The discriminator automatically selects the correct filter type based on column and operator. - union - … - `column` 'artifact.structuredOutputs[OutputID]', required — This is the column that will be queried in the call table. Structured Output Type columns are only to query on artifact.structuredOutputs[OutputID] column. - `operation` 'average' | 'count' | 'sum' | 'min' | 'max', required — This is the aggregation operation to perform on the column. When the column is a structured output type, the operation depends on the value of the structured output. If the structured output is a string or boolean, the operation must be "count". If the structured output is a number, the operation can be "average", "sum", "min", or "max". - `name` string — This is the name of the query. It will be used to label the query in the insight board on the UI. - `id` string, required — This is the unique identifier for the Insight. - `orgId` string, required — This is the unique identifier for the org that this Insight belongs to. - `createdAt` string, date-time, required — This is the ISO 8601 date-time string of when the Insight was created. - `updatedAt` string, date-time, required — This is the ISO 8601 date-time string of when the Insight was last updated. - `systemKey` string — Stable server-owned identifier for system-created insights. - TextInsight - `name` string — This is the name of the Insight. - `type` 'text', required — This is the type of the Insight. It is required to be `text` to create a text insight. - `formula` object — Formulas are mathematical expressions applied on the data returned by the queries to transform them before being used to create the insight. The formulas needs to be a valid mathematical expression, supported by MathJS - https://mathjs.org/docs/expressions/syntax.html A formula is created by using the query names as the variable. The formulas must contain at least one query name in the LiquidJS format {{query_name}} or {{['query name']}} which will be substituted with the query result. For example, if you have 2 queries, 'Was Booking Made' and 'Average Call Duration', you can create a formula like this: ``` {{['Query 1']}} / {{['Query 2']}} * 100 ``` ``` ({{[Query 1]}} * 10) + {{[Query 2]}} ``` This will take the You can also use the query names as the variable in the formula. - `timeRange` InsightTimeRange - `start` object — This is the start date for the time range. Should be a valid ISO 8601 date-time string or relative time string. If not provided, defaults to the 7 days ago. Relative time strings of the format "-{number}{unit}" are allowed. Valid units are: - d: days - h: hours - w: weeks - m: months - y: years - `end` object — This is the end date for the time range. Should be a valid ISO 8601 date-time string or relative time string. If not provided, defaults to now. Relative time strings of the format "-{number}{unit}" are allowed. Valid units are: - d: days - h: hours - w: weeks - m: months - y: years - `timezone` string — This is the timezone you want to set for the query. If not provided, defaults to UTC. - `queries` union[], required — These are the queries to run to generate the insight. For Text Insights, we only allow a single query, or require a formula if multiple queries are provided - union - JSONQueryOnCallTableWithStringTypeColumn - `type` 'vapiql-json', required — This is the type of query. Only allowed type is "vapiql-json". - `table` 'call', required — This is the table that will be queried. - `filters` union[] — This is the filters to apply to the insight. The discriminator automatically selects the correct filter type based on column and operator. - union - … - `column` 'id' | 'artifact.structuredOutputs[OutputID]', required — This is the column that will be queried in the selected table. Available columns depend on the selected table. String Type columns are columns where the rows store String data - `operation` 'count', required — This is the aggregation operation to perform on the column. When the column is a string type, the operation must be "count". - `name` string — This is the name of the query. It will be used to label the query in the insight board on the UI. - JSONQueryOnCallTableWithNumberTypeColumn - `type` 'vapiql-json', required — This is the type of query. Only allowed type is "vapiql-json". - `table` 'call', required — This is the table that will be queried. - `filters` union[] — This is the filters to apply to the insight. The discriminator automatically selects the correct filter type based on column and operator. - union - … - `column` 'cost' | 'duration' | 'averageModelLatency' | 'averageVoiceLatency' | 'averageTranscriberLatency' | 'averageTurnLatency' | 'averageEndpointingLatency' | 'artifact.structuredOutputs[OutputID]', required — This is the column that will be queried in the selected table. Available columns depend on the selected table. Number Type columns are columns where the rows store Number data - `operation` 'average' | 'sum' | 'min' | 'max', required — This is the aggregation operation to perform on the column. When the column is a number type, the operation must be one of the following: - average - sum - min - max - `name` string — This is the name of the query. It will be used to label the query in the insight board on the UI. - JSONQueryOnCallTableWithStructuredOutputColumn - `type` 'vapiql-json', required — This is the type of query. Only allowed type is "vapiql-json". - `table` 'call', required — This is the table that will be queried. - `filters` union[] — This is the filters to apply to the insight. The discriminator automatically selects the correct filter type based on column and operator. - union - … - `column` 'artifact.structuredOutputs[OutputID]', required — This is the column that will be queried in the call table. Structured Output Type columns are only to query on artifact.structuredOutputs[OutputID] column. - `operation` 'average' | 'count' | 'sum' | 'min' | 'max', required — This is the aggregation operation to perform on the column. When the column is a structured output type, the operation depends on the value of the structured output. If the structured output is a string or boolean, the operation must be "count". If the structured output is a number, the operation can be "average", "sum", "min", or "max". - `name` string — This is the name of the query. It will be used to label the query in the insight board on the UI. - `id` string, required — This is the unique identifier for the Insight. - `orgId` string, required — This is the unique identifier for the org that this Insight belongs to. - `createdAt` string, date-time, required — This is the ISO 8601 date-time string of when the Insight was created. - `updatedAt` string, date-time, required — This is the ISO 8601 date-time string of when the Insight was last updated. - `systemKey` string — Stable server-owned identifier for system-created insights. --- [API](https://skmtc.net/vapiai/apis/vapi-api.md) · [All operations](https://skmtc.net/vapiai/apis/vapi-api/llms.txt) · [OpenAPI document](https://skmtc-service-staging.skmtc.workers.dev/v1/apis/vapiai/vapi-api/versions/42e311e78e21/schema)