Roark
Let AI connect your sources for you
Skip the manual setup — run this in your project and the wizard auto-detects your databases and APIs and connects them to PostHog.
Connect Roark to PostHog to sync your data into the PostHog data warehouse for analysis and modeling.
Enter your Roark API key to sync your voice and chat AI observability data into the PostHog Data warehouse.
Generate an API key in your Roark dashboard. The key is a simple bearer token with read access to your organization's calls, chats, agents, metrics, issues, and simulations.
Configuration
| Option | Type | Required |
|---|---|---|
API key | password | Yes |
Linking Roark to PostHog
- Go to the Data pipeline page in PostHog
- Click New source and select Roark
- Fill in the required configuration fields
- Click Next, select the tables you want to sync, and then press Import
Supported tables
| Table | Description | Sync method | Incremental field | Primary key |
|---|---|---|---|---|
call | A monitored voice call between an AI agent and a customer, with Roark's analysis attached. | Full refresh | — | id |
chat | A monitored chat conversation between an AI agent and a customer. | Full refresh | — | id |
agent | A voice or chat AI agent monitored by Roark. | Full refresh | — | id |
agent_endpoint | A connection endpoint (phone number, websocket, etc.) an agent uses to place or receive calls. | Full refresh | — | id |
metric_definition | A definition of a metric Roark can compute over calls and chats (LLM judge, provider, threshold, formula, or pattern). | Full refresh | — | id |
metric_collection_job | A job that computes metric values over a set of calls or chats. | Full refresh | — | id |
issue | An issue Roark opened from an alert, trace failure, or manual report, grouping related evidence. | Full refresh | — | id |
simulation_scenario | A scripted scenario used to drive simulated calls against an agent. | Full refresh | — | id |
persona | A simulated customer persona (voice, accent, behavior) used to drive simulations. | Full refresh | — | id |
run_plan | A simulation run plan pairing scenarios, personas, and agent endpoints into a repeatable test suite. | Full refresh | — | id |
simulation_plan_job | A single execution of a simulation run plan. | Full refresh | — | simulationRunPlanJobId |
knowledge_base | A knowledge base document made available to agents. | Full refresh | — | id |