tvaras.ai

Data Output

Sync data from tvaras' internal database to your BigQuery warehouse. Use this for analytics, reporting, or integrating bot activity into your data pipeline.

What Gets Synced

Select which internal tables to export to BigQuery. Common tables include:

  • Tasks — task metadata, status, results
  • Task steps — step-by-step execution details
  • Task logs — event-level execution logs
  • Conversation events — chat messages and tool calls

Setup

  1. Go to Admin → Data Output
  2. Click Create Output
  3. Configure the destination:
    • GCP Project — BigQuery project ID
    • Dataset — target dataset name
    • Service Account Key — JSON key with BigQuery write access
  4. Select which tables to sync
  5. Set a schedule (cron expression) or trigger manually
  6. Optionally configure a webhook for completion notifications

Sync Behavior

  • Incremental — only new/updated rows are synced (using watermarks)
  • Batch loading — 2,000 rows per BigQuery load job
  • Type mapping — PostgreSQL types are automatically mapped to BigQuery types
  • Encrypted credentials — service account keys are encrypted at rest

Monitoring

In the admin panel, you can:

  • View sync execution logs (start time, duration, rows synced, errors)
  • Monitor incremental watermarks per table
  • Trigger manual syncs
  • Enable or disable outputs

Webhooks

Configure a webhook URL to receive notifications when syncs complete. The payload includes sync status, rows synced, and any errors. Requests are signed with HMAC-SHA256 using the webhook secret you provide.