What Governax records
Governax keeps a tamper-evident record of who changed access to what, across the tools your organisation already runs. When an auditor asks who granted an engineer production access in March, the answer is a query, not an archaeology project.
What Governax does#
Governax connects to your identity and infrastructure tools, watches them for access and configuration changes, and writes each change into an append-only ledger scoped to one entity. Every entry is chained to the one before it with a hash, so a later edit to the record is detectable rather than invisible.
Three properties are worth stating precisely, because they are the product:
- Append-only. The database physically refuses updates, deletes and truncates on the ledger table. This is enforced by Postgres triggers, not by application code that could be bypassed.
- Chained. Each event carries the hash of the previous event in its entity, so removing or altering one entry breaks every link after it.
- Verifiable on demand. You can re-walk an entity's whole chain at any time and get back a signed result.
The mechanics of all three are in The hash chain and append-only enforcement.
What Governax does not do#
Governax is a system of record, not a control plane. Every connector reads. Nothing Governax does can stop a change from happening, and nothing in the product grants access.
Related things that do not exist today, so you can plan around them: there is no policy engine, no compliance-framework mapping, no automated evidence collection, and no data export. There is also no public API. The only credential Governax accepts is a browser session cookie, so it cannot be called from a script, a scheduled job or another service, and there are no API keys to issue. If you need programmatic access, tell us what you are trying to do rather than which endpoint you want, because that shapes what gets built first.
How the pieces fit#
There are three nouns in Governax and it is worth learning them before anything else, because every URL and every permission decision is built from them.
Ledgers are per entity, not per organisation. Two entities never share a chain, and a connector belongs to exactly one entity. That is why product URLs look like /your-org/head-office/ledger. The reasoning is in Organisations and entities.
What is connected today#
Five connectors ship today. Three receive events pushed from the tool; two poll on a ten-minute interval, because the tool offers no practical push mechanism for an external service.
The difference matters when you are waiting for something to appear: a Slack change shows up in seconds, an AWS change can take up to ten minutes plus CloudTrail's own delivery lag.
Where to go next#
- New account? Start at Create your account.
- Already signed in and want events flowing? Quickstart is the shortest path.
- Evaluating the integrity claims? The hash chain and Verifying the ledger are the two pages to read.