Inspiration
While brainstorming ideas for the hackathon, we found ourselves doing what so many people do every day — struggling to communicate effectively with AI. We kept tweaking prompts, hoping for better results, only to get vague or off-target responses. That’s when it hit us: the problem isn’t the AI — it’s the prompt. _ So we stopped fighting the prompt and started fixing it. That’s how OptiPrompt was born — a smart solution to help people speak AI’s language with precision and purpose. _
What it does
OptiPrompt analyzes your input prompt and offers real-time suggestions to improve clarity, specificity, and relevance. Think of it as Grammarly, but for talking to large language models. Whether you're coding, writing, brainstorming, or querying — OptiPrompt makes sure you're getting the most out of your AI interactions. The goal? Better prompts. Better outputs. Less friction.
How we built it
- Streamlit: For the frontend — fast, interactive, and perfect for rapid prototyping with a clean UI.
- Google Gemini API: The brains behind the prompt optimization — analyzing input, enhancing clarity, and offering smarter rewrites.
- Python: Our go-to language for backend logic, API handling, and NLP-based preprocessing.
- OpenAI (ChatGPT): Used internally during development for code generation, debugging, and idea validation — basically our co-pilot.
Challenges we ran into
- Context awareness: Teaching the system to understand vague or poorly written prompts was tricky
- Keeping it concise: Optimizing prompts without bloating them required constant tweaking of our rewrite logic.
- Time constraints: It's a hackathon, so fitting a big vision into a tight time was its own wild ride.
Accomplishments that we're proud of
- We got the full pipeline working — from raw prompt to polished suggestion — in under 24 hours.
- Integrated Gemini smoothly and made it actually useful, not just flashy.
- The UI/UX is clean, intuitive, and just feels right.
- Most importantly? It works. We tested it on real prompts — and the output quality leveled up.
What we learned
- Prompt engineering is an art — and it’s still evolving.
- Gemini’s flexibility is powerful, but it takes finesse to harness.
What's next for OptiPrompt
- Use-case presets: Coding, content creation, research — each has different prompt needs. We’ll personalize optimizations.
- Chrome extension: Bring OptiPrompt anywhere — Notion, ChatGPT, Gemini, Docs, you name it.
Log in or sign up for Devpost to join the conversation.