How to create an Agent with its own documents as a Knowledge Source in under 5 minutes in Release0

3 minutes read

With Release0, you can create an Agent that uses its own documents (like PDFs, Word files, etc.) as a Knowledge Source in under 5 minutes. This is perfect for building AI assistants that can answer questions based on specific documents.

Creating a conversational AI Agent that understands your documents—like PDFs or Word files—has never been easier. With Release0, you can build and deploy an OpenAI-powered Agent that uses your custom documentation as a knowledge source, all in under 5 minutes.

🎥 Watch the full video tutorial


🚀 What You’ll Build

In this tutorial, we’ll guide you through creating an Agent that:

  • Accepts user input via a text field,
  • Uses OpenAI’s Assistant API,
  • Is trained with your business prompt and a PDF knowledge base,
  • Loops through questions to offer continuous support.

🛠 Step-by-Step: Create an Agent with Custom Document Knowledge

1. Start a New Agent

  • Go to the Agents section of your workspace and click Create Agent.
  • Choose Build from Scratch to start from a blank canvas.

2. Create a Welcome Message

  • Drag in a Text Dialog Block to greet the user.

3. Add a User Input Block

  • Add a User Input Block to capture questions. Set the input type to Text and save the response in a variable, e.g., UserMessage.

4. Connect to OpenAI

  • Add an OpenAI Integration Block to process the user’s question.
  • Create an API key at OpenAI platform
  • Paste the API key in the block.

5. Set Up the Assistant

  • Log into OpenAI and go to the Assistants section .
  • Create a new Assistant.
  • Add a custom prompt that trains the Agent about your business.
  • Upload PDFs or Word files under the Assistant’s “Knowledge” section. The Assistant will now respond using both your prompt and uploaded documents.

6. Connect the Assistant

Back in Release0:

  • Select your new Assistant from the OpenAI block.
  • Input: UserMessage
  • Output: Save response to a variable, e.g., AgentResponse.

7. Show the Agent’s Answer

  • Use another Text Dialog Block to display AgentResponse.

8. Loop for Continuous Conversation

  • Loop back with the User Input block to request the next question from the user.

🧠 Why This Is Powerful?

By combining OpenAI Assistants with Release0’s no-code builder, you can:

  • Deploy product-aware AI Agents instantly,
  • Answer user questions with reliable knowledge sources,
  • Reduce support load without sacrificing quality .

🌍 Share It Anywhere

Once published, your Agent can be:

Share this article

0