> ## Documentation Index
> Fetch the complete documentation index at: https://docs.synheart.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Audit log

> Who changed what in Synapse, and when. What's recorded, what never is, filtering by action, the detail on each event, and exporting it as a CSV

The **Audit log**, under **Admin → Trust** at [synapse.synheart.ai](https://synapse.synheart.ai), records who did what, when, and to what, across your organization's Synapse. Nothing in it can be edited or deleted.

Synapse Admins and the organization owner, with a seat, can open it. Anyone else is pointed to their own **Privacy & consent** page.

## What's recorded

Governance actions, as they happen, grouped by what they're about:

| Group | For example |
| - | - |
| **Connectors** | Connecting, updating or disconnecting a source, and disconnecting with its data purged |
| **Knowledge sources** | Adding, keeping or stopping a source |
| **Personal accounts** | Linking, unlinking or verifying an account |
| **MCP tokens** | Issuing and revoking a token |
| **Exports** | Downloading a CSV or JSON export, and creating, changing, testing or deleting a webhook |
| **Goals & alerts**, **Digests** | Creating, changing and deleting goals and alerts, changing the weekly digest, and sending one |
| **Departments** | Creating, renaming and deleting a department, and assigning or removing members |
| **Agent rules** | Saving a new version of the team rules |
| **Privacy**, **Settings** | Changing the cohort floor, team human-state views, team sharing or the reporting time zone |
| **Synaptic Mesh** | Restricting and un-restricting a document |
| **AI rollouts**, **AI adoption** | Recording a rollout, and matching AI tool users by work email |
| **Synapse Life** | Connecting and disconnecting Life |
| **Offboarding**, **Memory** | Erasing a member's or the organization's data, and forgetting a memory |
| **Audit log** | Exporting the log itself |

Goals and alerts that Synapse retired itself, when the AI Effectiveness Index stopped being a single score, are marked **Retired**.

## What's never recorded

The log records actions, never content. It never holds:

* a credential, token, API key or secret, though it may name one, such as a token's label;
* message text, documents, prompts or code;
* anyone's human state;
* your own consent choices, such as sharing your numbers with your team. Those are kept as proof of consent but never shown here. See [Team views and privacy](https://docs.synheart.ai/synapse/team-views.md).

## Reading an event

Each row shows when it was recorded, who did it and what they did. Open a row for its details:

* **Actor**: the person, or **System** for something Synapse did on its own. A person whose data was erased shows as **Erased user**.
* **Target**: what it was done to, such as a department, a connector or a goal.
* **Recorded**: the exact time.
* **Action**: the action's code, such as `department.created`.
* Any detail that applies, such as the metric and threshold of an alert, the channel of a digest, or the number of rows in an export.

## Filtering and paging

**Filter by action** shows one kind of event. Choose how many to show **Per page** and move between pages. **Refresh** loads anything new.

## Exporting

**Export CSV** downloads every event matching the current filter, not just the page on screen, up to 20,000 rows. A bigger match is refused rather than cut short, so a file you download is always complete for its filter. Narrow the filter and try again.

Every export is itself recorded in the log.

## Who can do what

| Action | Who |
| - | - |
| Read and export the audit log | Synapse Admins, and the organization owner with a seat |


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.