/Respell-Call-Streamlit-Demo

A Streamlit app that uses Respell AI's API to allow you to make calls using AI.

Primary LanguagePythonMIT LicenseMIT

Respell Call Streamlit App 📞

Streamlit App

Ring, Ring, who's calling? ☎️

Check out how I gave myself a ring using this app! 📞🤙

Conversation Catalyst

Welcome to the Respell Call Streamlit App, your gateway to initiating calls with a sprinkle of awesomeness. Let's cut to the chase and dive right in, shall we?

What's this all about? 🤔

This Streamlit app is your virtual assistant, making calls to the phone number you provide and carrying out the objective you set.

Features 🫠

  • User-Friendly Interface: Streamlit UI which makes building Python apps easy-peasy lemon squeezy!

  • Secure API Calls: Utilizes the Respell API for secure and efficient calls.

  • Content Moderation: Uses Lakera AI for real-time prompt moderation to keep the conversation clean and respectful.

  • Interactive Chat Transcript: You get a neat transcript of your call, so you can analyze all the little details afterwards.

How to Get Started 🎉

  1. Clone the repository:

    git clone https://github.com/tonykipkemboi/Respell-Call-Streamlit-Demo.git

  2. Install requirements:

    pip install -r requirements.txt

  3. Add your API keys to a secrets.toml file in the root directory of this project. Streamlit will pick it up automatically when you run the app.

    Here's how your secrets.toml file should look like:

    [default]
RESPELL_API = "your-respell-api-key-goes-here"
SPELL_ID = "your-spell-id-goes-here"
SPELL_VERSION = "your-spell-version-goes-here"
LAKERA_API_KEY = "your-lakera-api-key-goes-here"

  4. Run the App - In your terminal, navigate to the folder where the app is located and type:

    streamlit run streamlit_app.py

  5. Enjoy! - Play around with the app, and don't forget to wear a smile!

Behind the Scenes 🧙🏽‍♂️

  • parse_transcript Function: This function takes the transcript text, splits it into separate messages, and filters out the unwanted substrings, giving you a clean and structured output.

  • Moderation with Lakera: Before making a call, the objective (prompt) text is sent to Lakera for moderation. If it contains hate speech or sexual content, the call won't proceed. So, keep it classy!

  • Session State: We use the session state to hold onto your messages throughout your session.

This Streamlit app is powered by 🔋

Conversation Catalyst

Conversation Catalyst