​
General Overview

​
Introduction

MultiBot Agents provide a flexible framework for creating and managing custom AI assistants powered by various model providers. This feature allows users to build specialized AI agents with unique configurations, capabilities, and integrations—without requiring extensive coding knowledge.

MultiBot Agents function similarly to OpenAI’s Assistants API and ChatGPT’s GPTs but offer broader model support and a no-code implementation, making it easy to deploy AI-driven solutions for different tasks.

This guide provides a comprehensive walkthrough of setting up, configuring, and utilizing MultiBot Agents effectively.

​
Getting Started with MultiBot Agents

​
Creating a New Agent

To create a new agent:

  1. Open the MultiBot Agents Panel

    • Navigate to the “Agents” section from the endpoint menu.

    • Click on the Agent Builder in the Side Panel.

  2. Fill Out Agent Details

    • Avatar: Upload a custom avatar for personalization.

    • Name: Choose a distinctive name for your agent.

    • Description: (Optional) Add a brief summary of your agent’s purpose.

    • Instructions: Define system instructions to shape the agent’s behavior.

    • Model: Select from available AI providers and models.

  3. Select an Existing Agent

    • Agents can be selected from the top dropdown menu in the Side Panel.

    • You can also mention an agent in the chat using ”@” in the input field.

​
Model Configuration

The Model Parameters Interface allows fine-tuning of your agent’s responses:

  • Temperature (0-1): Adjust response creativity.

  • Max Context Tokens: Set the memory limit for the conversation.

  • Max Output Tokens: Define the maximum response length.

  • Provider-Specific Settings: Customize parameters based on the chosen AI model.

​
Agent Capabilities

​
1. Code Interpreter

When enabled, the Code Interpreter allows your agent to:

  • Execute code in multiple programming languages, including:Python, JavaScript, TypeScript, Go, C, C++, Java, PHP, Rust, and Fortran.

  • Process files securely through the MultiBot Code Interpreter API.

  • Run code without requiring local setup, configuration, or sandbox deployment.

  • Handle file uploads and downloads seamlessly.

The File Search feature enables:

  • Retrieval-Augmented Generation (RAG): AI can retrieve relevant information from uploaded documents.

  • Semantic Search: AI understands document content contextually.

  • Context-Aware Responses: AI can reference uploaded files in conversations.

  • File Attachment Support: Attach documents at both the agent and chat thread levels.

​
Tools & Integrations

MultiBot Agents can be enhanced with various built-in tools:

  • DALL-E 3: Generate images from text descriptions.

  • Tavily Search: Advanced search API with diverse data sources.

  • Calculator: Perform mathematical calculations.

  • Google Search: Retrieve real-time web search results.

  • Stable Diffusion: Generate images from text prompts.

  • Azure AI Search: Retrieve structured information.

  • Traversaal: A robust search API for LLM Agents.

  • Wolfram Alpha: Perform complex computations and mathematical reasoning.

​
Actions & Custom API Integrations

With Actions, you can extend your agent’s functionality by adding custom tools from OpenAPI specifications.

​
Adding Custom Actions

  1. Click the “Create Action” button in the Agent Builder.

  2. Enter the OpenAPI spec URL to generate a new tool.

  3. Save and assign the action to your agent.

Note: Individual domains can be whitelisted for agent actions.

​
File Management

MultiBot Agents support three types of file uploads:

  1. Image Uploads: For AI-based visual content processing.

  2. File Search Uploads: Upload documents for RAG capabilities.

  3. Code Interpreter Uploads: Upload files for code execution.

Files can be attached either: ✅ Directly to the agent configuration ✅ Within individual chat threads

​
Best Practices

To ensure the best experience with MultiBot Agents:

✅ Provide clear, specific instructions for your agent. ✅ Carefully select the necessary tools for your use case. ✅ Organize files properly across the three upload categories. ✅ Review sharing permissions before making an agent public. ✅ Test your agent thoroughly before deploying it.

​
Recap: How to Create & Use MultiBot Agents

  1. Select “Agents” from the endpoint dropdown menu.

  2. Open the Agent Builder panel.

  3. Fill in the required details (name, avatar, model, instructions).

  4. Configure capabilities (Code Interpreter, File Search, Tools).

  5. Add necessary tools and files.

  6. Set sharing permissions (if needed).

  7. Create & start using your agent! 🎉

​
Conclusion

MultiBot Agents provide a powerful, no-code solution for building AI-driven assistants with customizable capabilities. Whether you’re creating chatbots, automation tools, or research assistants, MultiBot Agents offer seamless integration with your preferred AI models and services.

By following this guide, you can quickly set up, deploy, and scale AI agents tailored to your specific needs.