README.md

Event Finder with Reka Research

This demo shows how to use the with the Reka Research with Reka API to search the web and return structured output using response_format.

It's built with Streamlit.

Features

• Uses response_format to get structured event data (title, date, URL)
• Uses research["web_search"] to control which domains are searched
• Displays reasoning steps during streaming
• Shows results in styled event cards based on structured output

📚 Learn more in our documentation:

• Strucutred Output
• Web Search

How it works

1. User enters a query like:

   Find AI conferences in Tokyo next month

2. User can choose to:

   • Allow all web results
   • Limit results to specific domains (e.g. eventbrite.com)
   • Block results from specific domains

3. The app sends a streaming request to the Reka API with:

   • A JSON schema as response_format
   • A research config as research

4. The model searches the web and returns structured results like:

   {
     "events": [
       {
         "title": "AI Expo Tokyo",
         "date": "2025-07-10",
         "url": "https://aiexpo.jp"
       }
     ]
   }

5. The app renders this as UI elements in the browser.

Running the app

1. Make sure you have Python, Streamlit and OpenAI SDK installed.

   pip install streamlit openai

2. Set your Reka API key in the environment:

   export REKA_API_KEY=your-api-key

   Sign up for a Reka API key at Reka Platform if you don’t already have one.

3. Run the app:

   streamlit run app.py

---

File structure

• app.py: Main Streamlit app
• README.md: This guide

Notes

• You can change the schema to get other structured formats, like job listings or product data.
• This app uses reka-flash-research with streaming enabled.