What data does Callmine fetch from Gong?

TL;DR

For each call in a date range you pick, Callmine fetches metadata, participants, transcript text, and any HubSpot deal context Gong has linked to the call. Retention runs from 30 days on Trial and Free up to unlimited on Scale.

What we fetch

Per analyzed call, Callmine reads:

• ·Call metadata — start time, duration, direction, title, participants, rep identity — via /v2/calls/extensive.
• ·Transcript text — the speaker-attributed conversation log.
• ·HubSpot deal context, when present — deal stage, amount, owner, company, and any custom properties Gong has logged against the call. This is read from the same /v2/calls/extensive payload; no separate HubSpot credential is required for the call-embedded fields.

We do not fetch audio, video, scorecards, or notes. Those stay in Gong.

Where it lives

Call metadata, report rows, schedules, and audit entries are stored in Postgres on Cloud SQL. Larger payloads — raw call records, full transcripts, completed report JSON — are written to a Google Cloud Storage bucket. Both run in Google Cloud's us-central1 region.

Encryption

Everything we store is encrypted at rest by Google Cloud's infrastructure. On top of that, your workspace's Gong credential is encrypted application-side with a symmetric key held in our secret manager — so even with full database access, an attacker cannot read your access key or refresh token without the application key.

In transit, every hop runs over HTTPS/TLS: Callmine to Gong, your browser to Callmine, and Callmine to the LLM providers that run the analysis prompt.

Retention by plan

Data is never deleted from our storage when retention expires — older reports are simply hidden from the UI until you upgrade. If you want a permanent deletion (for example, before disconnecting), email hello@callmine.ai and we'll process it manually.