Skip to main content

​
1. Select your model provider

When you log in to Devic, the first step is to choose the model provider that will run your agents and assistants. In the top bar, you’ll find the model selector, with support for the following providers:
  • OpenAI
  • DeepSeek
  • Anthropic
  • XAi
  • Kimi-K2

Provider configuration

Click the settings icon next to the provider name to authenticate using your API key.

​
Provider authentication

Before creating or running agents, you must authenticate your account using the API Key of the selected model.
This step is mandatory: the platform will not allow you to proceed until authentication is completed successfully. In the configuration window, enter:
  • Project Name: an identifier for your project or environment.
  • Project API Key: the API key issued by the model provider.
Provider authentication To obtain your API key, refer to the official guides: Once authenticated, the model becomes available for use in agents, assistants, and MCPs.

​
2. Dashboard

The Dashboard is the central control point of Devic.
Here you can monitor agent performance and the status of executions in real time.

Executions pending approval

Displays tasks paused under the Human-in-the-Loop workflow that require manual approval before continuing.

Recent executions

List of completed or ongoing executions, including date, status, and the agent involved.

Executions in progress

Shows which agents are active and the tasks currently being executed.

Upcoming executions

Schedule of agents configured for automatic execution.

Performance charts

Real-time metrics for agents executed in the last 12 hours.
Main dashboard

​
3. Main areas

The left sidebar groups all functional sections of Devic.
From here you can access each main module to configure, monitor, and operate your AI-native environment. Main sidebar

Assistants

Conversational interfaces that respond in real time. You can connect them to your data or integrate them into your products.

Agents

Autonomous entities that execute business workflows, invoke tools (MCPs), and manage complex tasks.

Databases

Vector databases with integrated semantic search, available for your agents and MCPs.

MCPs

Tools and connectors that extend your agents’ capabilities using the Model Context Protocol standard.

Next step

Create your first MCP and start giving your agents the tools they need.