# AI agents & automations

URL: https://openship.io/features/ai-agents

Investigate deployments and run scoped operational tasks with your own agent.

Connect Codex or Claude Code, create specialist agents and schedule approved automations. Work from the same application context and resource permissions used by the platform.

The assistant can start from deployment failures, app issues and health reports. Review its requested access, connect your own agent account or supported API configuration, and choose where the agent runs.

Agents and automations use the approved workspace permissions. Desktop can use a local agent; self-hosted installations use their host, and Cloud can use an authorized paid instance or connected server.

## Contextual investigation

Start from the app, deployment or incident you are reviewing.

## Your agent connection

Use supported native agents with your chosen account or API configuration.

## Scheduled work

Create scoped specialists and recurring automations with run history.

[Read the guide](https://openship.io/docs/dashboard/assistant)
