Agenta: build AI coworkers and scheduled automations

Learn how Agenta uses company context and editable instructions for team agents, with examples of SEO research and daily issue reports sent to Slack.

Player not loading? Watch on YouTube

This launch walkthrough introduces Agenta, an open source workspace for building AI coworkers that a team can share. The presenter describes agents for marketing, product and engineering, then demonstrates requests for research and recurring work. The workspace supports cloud use or a self-hosted deployment, though the video does not explain installation or hardware requirements.

The first example asks an AI agent to research questions on Reddit, check keyword opportunities in Ahrefs and produce an SEO article plan. The presenter says the agent can continue working in the background after he closes his laptop. He also says it has access to previous research, positioning decisions and other company context.

A second example addresses writing preferences. The presenter asks the agent to remember to avoid em dashes and says it updates its configuration in response. Users can also inspect and edit the instructions themselves. For recurring work, he asks a product agent to review new issues each morning, identify high priorities and send a report to Slack.

The presenter claims broad model compatibility and support for existing AI subscriptions. The walkthrough does not establish which models or subscriptions are supported, or demonstrate offline execution. Its focus is agent behavior and automation rather than deployment setup.