SpeakBuddyHF26: Your AI English Speaking Partner

This is a submission for the Hacktoberfest Weekend Challenge: Build for a Friend

What I Built

I built SpeakBuddyHF26, an AI English Speaking Partner for a friend who wants to improve their English speaking confidence without the fear of being judged for making mistakes.

SpeakBuddy creates a friendly and patient environment where users can practice realistic English conversations using their microphone.

Users can:

• Choose Beginner, Intermediate, or Advanced speaking levels
• Practice Daily Conversations
• Practice Job Interviews
• Practice College Conversations
• Practice Travel situations
• Practice a 60-second personal introduction
• Speak naturally using their microphone
• Continue an interactive conversation with the AI
• Track their speaking practice and progress

The goal is to make English practice feel like talking to a supportive friend rather than attending a classroom lesson.

Demo

🎥 2-minute demo:

https://youtu.be/hR1RFzRCrZk?si=szuy9MlMgs0s4TiC

The demo shows the complete experience, including selecting a scenario, speaking with SpeakBuddy, receiving AI responses, and viewing the application.

Code

💻 GitHub Repository:

https://github.com/NavyaPachigolla/SpeakBuddy

How I Built It

SpeakBuddy is built around Llama 3 8B running locally through Ollama.

The basic flow is:

User speaks → Speech is transcribed → Llama 3 processes the conversation → AI responds → User continues the conversation

The project includes:

• Llama 3 8B
• Ollama
• Local AI inference
• Voice/microphone interaction
• Real-world conversation scenarios
• Multiple speaking levels
• Progress tracking
• Full-stack frontend and backend

Why Does Open Innovation Matter?

Open innovation is important for SpeakBuddy because the application is designed to provide a private and comfortable environment for English speaking practice.

I use Llama 3 8B through Ollama, allowing the core AI conversation experience to run locally instead of depending on a closed proprietary AI API.

This gives the project more flexibility and helps keep the user's practice experience private.

For someone practicing English, they may talk about their education, career goals, personal experiences, and everyday life. Local AI provides a useful way to practice without requiring those conversations to be sent to a third-party AI service.

My Agent Session

I used an AI-assisted development workflow to build and improve SpeakBuddyHF26.

What I Learned

Building SpeakBuddy taught me that an AI application does not need to be complicated to be useful. The most important part was designing the experience around a real person's problem.

I focused on making the conversations simple, encouraging, and practical so that the user can keep practicing without feeling judged.

Final Thoughts

SpeakBuddyHF26 was built to make English speaking practice more comfortable, private, and accessible.

Instead of worrying about making mistakes, the user can simply speak, practice, learn, and try again.

Thank you for checking out my Hacktoberfest 2026 project!

Comments

[Navya Pachigolla]

Thanks for checking out SpeakBuddyHF26! I built it to make English speaking practice more comfortable, private, and friendly. I’d love to hear your feedback.