Looking for a Good Analytics Tool for Voiceflow AI Agents
Does anyone have a recommendation for an analytics tool for my AI agents built with Voiceflow?
I was using Chatdash before, but they made it to expensive.
1 Reply
Here are the main ways Voiceflow builders are tracking agent performance today (no extra licences required):
1. Built-in Analytics Dashboard
• Open your agent and click the Analytics tab (shortcut 4).
• Out-of-the-box charts for Credits Usage, Total Interactions, Latency, Unique Users, Intents/Functions/Prompt breakdowns, etc.
• Data is available for any traffic that happened on/after Feb 9 2025. If that covers your date range this is the quickest (and free) replacement for Chatdash. 2. Analytics API (for your own BI tool)
• REST endpoint:
• Pull raw usage / metric data into Looker Studio, Power BI, BigQuery, Mixpanel, etc., and build whatever dashboards you like.
• Also exposes a Transcripts API if you want conversation-level detail. 3. Advanced Analytics Proxy recipe (real-time trace capture)
If you need fine-grained data such as every trace coming back from the Dialog Manager (e.g. to feed Datadog or a custom DB), Voiceflow published an open-source proxy you can self-host:
GitHub: https://github.com/voiceflow-gallagan/vf-analytics-proxy 4. CLI quick checks
Run
• Open your agent and click the Analytics tab (shortcut 4).
• Out-of-the-box charts for Credits Usage, Total Interactions, Latency, Unique Users, Intents/Functions/Prompt breakdowns, etc.
• Data is available for any traffic that happened on/after Feb 9 2025. If that covers your date range this is the quickest (and free) replacement for Chatdash. 2. Analytics API (for your own BI tool)
• REST endpoint:
GET /v1/analytics/query → docs: https://developer.voiceflow.com/reference/querypubliccontroller_queryusage• Pull raw usage / metric data into Looker Studio, Power BI, BigQuery, Mixpanel, etc., and build whatever dashboards you like.
• Also exposes a Transcripts API if you want conversation-level detail. 3. Advanced Analytics Proxy recipe (real-time trace capture)
If you need fine-grained data such as every trace coming back from the Dialog Manager (e.g. to feed Datadog or a custom DB), Voiceflow published an open-source proxy you can self-host:
GitHub: https://github.com/voiceflow-gallagan/vf-analytics-proxy 4. CLI quick checks
Run
voiceflow analytics in the Voiceflow CLI to fetch high-level metrics right from your terminal or CI pipeline.
In short: try the new in-product Analytics tab first—most teams find it covers their reporting needs without extra cost. If you need more flexibility, the Analytics API or the proxy recipe lets you stream data into whichever analytics stack you prefer.