Skip to main content
Everything Corbado Observe reconstructs is available outside the management console. You can query it live, export it in open formats and load it into the platforms you already run. You receive the same data model the console uses: classified journeys with their subflows, decisions and errors, not only raw events.
The Observe data model of flows, subflows and events feeds the management console for dashboards and investigation, the Observe API with scoped keys, the Observe CLI for scripts and AI agents, daily table exports as Parquet files and time-series exports as CSV or Parquet. All five use the Observe data model. Match reporting intervals and filters when comparing KPIs, and account for snapshot freshness.The Observe data model of flows, subflows and events feeds the management console for dashboards and investigation, the Observe API with scoped keys, the Observe CLI for scripts and AI agents, daily table exports as Parquet files and time-series exports as CSV or Parquet. All five use the Observe data model. Match reporting intervals and filters when comparing KPIs, and account for snapshot freshness.

​
1. Ways to use your data outside the console

The API and CLI reflect the current state of your project. Table exports are written once per day for the previous day, so they suit history and joins with your own data rather than live monitoring. For enterprise delivery requirements, discuss custom exports with Corbado.

​
2. What the data contains

  • Classified journeys: flows such as login, sign-up and recovery, their subflows such as passkey login or email OTP, the options users were offered and the errors they hit. See the data model.
  • Raw and low-level events: the events your integration sent and interaction signals such as focus, input and paste, for your own analysis.
  • Environment data: browser, operating system, device and WebAuthn client capabilities per session.
  • KPI series: the same pre-calculated series the console reads. Match the reporting interval, dimensions and filters when comparing numbers.

​
3. Security and privacy

  • Scoped keys: every API key carries explicit permissions. Analysis and exports need read permissions only. Keys can be limited to IP addresses or ranges and can expire.
  • Pseudonymous identifiers: Observe stores the user identifier your integration sends. Send a pseudonymous value, for example a keyed hash, and compute the same value in your own systems to join the data without exposing the raw ID.
  • No input values: the Web SDK records how a field was filled, for example typed, pasted or autofilled, but not what was entered. From Web SDK version 0.16.7, URL query values are masked by default.
  • Per-user requests: export or delete all Observe data of a user through the API. See per-user export and deletion.
Hosting options, including dedicated instances, are described under projects, environments and hosting.

​
4. Compatible platforms

Exports use Apache Parquet and CSV, and the API returns JSON. Observe data therefore works with any platform that reads these formats, from lakehouses and warehouses to BI and monitoring tools. Connect your data platform shows the loading pattern and notes for common platforms.

Data model

Tables, keys, time columns and schema changes.

Observe API

Authentication, permissions, endpoints and limits.

Exports

Table exports, time-series exports and per-user requests.

Connect your data platform

Load exports safely into your warehouse or lakehouse.