Screenshots

Agents run CLI commands and Python scripts inside a sandboxed environment. The sandbox restricts access to system files and enforces proper filesystem isolation, preventing privilege escalation.

Agents can search the web in real time using SearXNG. Here an agent researches gold price trends and presents a structured summary with a data table and key takeaways.

Agents can delegate tasks to other agents. In this example, the Dark Matter agent delegates a question to the Oracle agent, which returns its response inline in the same conversation.

Agents learn from the current conversation and carry that knowledge into future interactions. Here the agent learns the user’s name and preferences, storing them in memory for later use.

Agents learn from the current conversation and carry that knowledge into future interactions. Here the agent learns the user's name and preferences, storing them in memory for later use.

The Agents tab lists all available agents on the instance. Each agent has its own identity, personality, and set of enabled tools. Green indicators show which agents are active.

The Tasks tab shows background tasks that agents are working on. Agents can be assigned tasks that run autonomously, with status tracking and results displayed when complete.

Create an account with a username, name, email, and password. The first account created on a new instance becomes the first user.