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

# Security and data handling

> How Fieldwork protects your data

## Where your data lives

Fieldwork runs on AWS. Application data lives in encrypted managed PostgreSQL; files and datasets in encrypted S3 storage. All traffic is TLS in transit. Production infrastructure is isolated in its own AWS account with private networking — databases and workloads are not reachable from the internet.

## Organization isolation

Your organization is the trust boundary. Conversations, connectors, files, tables, the knowledge base, automations, and skills are all organization-scoped, with per-asset access controls on top. Data reads are authorized against live access-control state at the moment of access — revoking access takes effect immediately.

Shared conversations never widen access: a teammate's turns in your conversation run under **their** identity and permissions, not yours.

## The workspace sandbox

The agent's code execution happens in isolated, per-conversation containers, provisioned fresh and torn down after use. Workspaces reach your organization's data only through brokered, authorized channels — there is no standing credential inside the sandbox.

## Credentials and connectors

* Connector authorizations use each provider's OAuth flow; Fieldwork holds only the scopes you approved
* Stored tokens are encrypted at rest
* Secrets are entered through write-only forms — they never transit chat, and the agent can never read them back
* Connector actions run under the connected account's real permissions; approval cards gate anything that changes external systems

## AI models

Fieldwork uses frontier models from Anthropic, OpenAI, and Google, accessed through enterprise API agreements. Your data is not used to train these models. Prompts are assembled server-side from your organization's data at request time — model providers receive conversation context, not standing access to your systems.

## Our security program

Fieldwork maintains a continuously monitored security program — see our **[Trust Center](https://trust.getfieldwork.ai)** for live control status. Service status and incident history are published at **[status.getfieldwork.ai](https://status.getfieldwork.ai)**. Our sub-processors are listed at [getfieldwork.ai/subprocessors](https://getfieldwork.ai/subprocessors), and our privacy policy at [getfieldwork.ai/privacy](https://getfieldwork.ai/privacy).

To report a security concern: **[security@getfieldwork.ai](mailto:security@getfieldwork.ai)**.