Quickstart

Connect your agent and run your first evaluation in minutes.

1. Create an account

   Go to arkdock.arklex.ai and sign up with your work email. Click the verification link in the email you receive to activate your account.

   If your team already has an organization, ask an Owner to invite you from Settings > Members instead.

2. Connect an agent

   Navigate to Agents and click Connect Agent. Enter a name, select the API type (Chat Completions for most agents), and paste your endpoint URL. Add any auth headers your endpoint requires, then click Test Connection. Save when the connection is confirmed.

   
3. Create scenarios

   Navigate to Scenarios and click Add Scenarios. Choose Generate to have Arkdock create scenarios from a description of your agent and its users, or choose Import to upload an existing file. Save the scenarios to a group.

4. Run a simulation

   Navigate to Simulations and click New Simulation. Select your agent and the scenarios you just created, then click Run Simulation. Arkdock runs each scenario as a full multi-turn conversation and shows the transcripts when done.

   
5. Run an evaluation

   Once the simulation shows Completed, navigate to Evaluations and click New Evaluation. Select the simulation, choose your metrics (Goal Completion is always included), and click Run Evaluation.

6. Review results

   Open the evaluation to see per-metric scores, detected behavioral errors, and scored conversation transcripts. Click any conversation row to read the full transcript and the LLM judge’s reasoning for each turn.

   

---

Next steps

• Customize metrics to score domain-specific behaviors.
• Upload knowledge documents to ground simulated users in your product content.
• Invite teammates to annotate and calibrate evaluation results together.