Your first conversation

Go from a new workspace to a voice conversation in the dashboard.

1. Open your workspace

Sign in at app.evacompany.co with Google. EVA opens your workspace automatically when you have one. If you belong to several, choose the workspace you want to use.

The selector at the top left lets you switch workspaces or create another. Your account menu is at the bottom left.

2. Create an agent

Open Build → Agents and create an agent. Give it a recognizable name. Choose the EVA engine for a configurable speech pipeline, or Cartesia to connect an existing hosted agent.

For an assisted start, open EVA Asistant, describe your business, and provide your public website. Review the draft it creates before publishing.

3. Set its purpose

Tell the agent what it should accomplish, the facts it can use, and when to ask for help. Select speech recognition, a language model, and a voice from the configuration menus.

For example:

You help guests ask about our hotel. Answer from the business information provided. Ask one question at a time. Do not claim that a booking is confirmed. If availability is unknown, offer a staff follow-up.

4. Publish and test

Save the draft, then publish a version. Start a test conversation from the agent page, allow microphone access, and use headphones to reduce feedback. Test an ordinary question, a short interruption, and a question the agent cannot answer.

Publishing creates an immutable version. Draft edits do not change a version already used by a call.

5. Review the result

Stop the call, then open Data → Call History. The transcript, recording status, and analysis are separate results. Recording and analysis may finish after the conversation ends.

Next: configure your agent, connect phone calls, or use the API.