The Studio Build Path offers a low-code, GUI-driven experience for designing, configuring, and deploying AI agents in Lyzr. It guides you through each stage using intuitive interfaces and wizards, enabling rapid prototyping and iteration without writing a single line of code.

​
Key Steps

  1. Project Initialization

    • In the Studio, click Build Agent.

  2. Data Connector Setup

    • Navigate to Data Connectors in the Project dashboard.
    • Click Add Connector, choose a source (databases, file uploads), and fill in connection details.
    • Test connectivity to ensure credentials and network settings are correct.

  3. Knowledge Base Configuration

    • Go to Knowledge Bases and click Create KB.
    • Provide a name, description, and select the vector database backend.
    • Upload documents (PDF, DOCX, CSV, JSON) or link to external sources.
    • Let the Studio auto-select parsers, chunk content, generate embeddings, and index data.
    • Configure Responsible AI settings (facts, reflection, groundedness, context relevance) under Safety & Responsibility.
    • Attach custom tools by browsing the Tools section or adding new ones via Create Tool.

  4. Testing & Iteration

    • Open the Simulator to enter test prompts and observe responses in real time.
    • Adjust retrieval parameters (number of chunks, score threshold) and safety settings to refine output.
    • Monitor logs and analytics in the Insights tab to identify errors and optimize performance.

  5. Deployment & Management

    • When ready, click Publish to deploy your agent.
    • Manage agent versions, set environment variables, and assign access controls in Settings.
    • Track usage, error rates, and user feedback through the Dashboard, iterating as needed.