Inspiration
Language barriers limit global communication, making interactions difficult for travelers, remote teams, and businesses. We wanted to create a seamless way for people to chat in different languages without friction.
What it does
Ziadico is an AI-powered chatbot that provides real-time message translation, enabling users to communicate effortlessly across languages. Currently integrated with Telegram, it helps individuals and teams break language barriers instantly.
How we built it
We developed Ziadico using:
- Programming Language: Node.js (JavaScript)
- Deployment: Railway
- API: Telegram API and Gemini API
- Datebase: Mongodb
- LLM: Gemini
- Integration & Use: Built for and currently available on Telegram
Due to cost considerations, we launched on Telegram first, ensuring accessibility and ease of use.
Challenges we ran into
- Balancing translation accuracy with response speed.
- Managing API rate limits and cost constraints.
- Ensuring a seamless user experience within Telegram’s interface.
- Issues finding free deployment services, leading us to use free trials for paid services to host and run the chatbot.
Accomplishments that we're proud of
- Successfully deploying a functional AI chatbot with real-time translation.
- Providing a smooth user experience on Telegram.
- Receiving positive feedback from early testers on translation accuracy.
What we learned
- Cost-effective deployment is critical for scaling AI-based solutions.
- Users need a highly intuitive experience to trust and adopt AI chatbots.
- Real-time translation needs continuous optimization for tone and context.
What's next for Ziadico
We plan to expand beyond Telegram by:
- Building a standalone app for a more personalized experience.
- Integrating with workplace platforms like Slack, WhatsApp, and Teams for B2B & B2C commercialization.
- Enhancing AI capabilities to support industry-specific language models.
- Shipping the following features:
- Voice translation & text-to-speech to enable seamless spoken conversations.
- AI understanding slang and informal speech for more natural interactions.
- Features are prioritized using the Jobs to Be Done (JTBD) and RICE frameworks, based on users feedback and suggestions.
Log in or sign up for Devpost to join the conversation.