This content originally appeared on DEV Community and was authored by ujjavala
Generative AI has quickly become part of everything—writing, coding, art, music. It’s exciting, sometimes overwhelming, and undeniably powerful. These models automate tasks, accelerate workflows, and unlock new kinds of creativity.
But beyond the buzz and breakthroughs, there’s a quieter side that’s just as important: how AI can be used to support personal growth, encourage reflection, and respond with empathy. It’s not just about what the AI can do—it’s about how we design the interaction.
That’s where prompting comes in. Prompts aren’t just technical instructions; they shape tone, intent, and emotional texture. A well-crafted prompt can turn a model from a machine into something that feels more like a guide, a coach, or even a companion.
This blog shares a small, personal exploration of that idea, a web-based app built with open-source tools running locally, designed to support speech practice through gentle, AI-driven interaction. It focuses on encouragement rather than correction, and prioritizes privacy, calm, and kindness.
To my surprise, using the app felt remarkably close to moments in real speech therapy—focused, reflective, and safe. While it’s not a substitute for a human therapist, the AI’s steady, judgment-free responses created space to practice with ease. Whether suggesting simpler alternatives, offering pacing advice, or just replying with care, it reminded me that how we prompt—and why—matters just as much as what we build.
Starting with a Feeling, Not a Roadmap
This app started with a simple curiosity:
What if AI could help someone practice everyday conversations—like ordering coffee or making a phone call, without feeling judged?
Not trying to fix anything, just encouraging. Not scoring or assessing, just guiding.
From that, a simple flow took shape:
- Type a sentence you want to practice
- Get a few easier, friendlier ways to say it
- See gentle tips on pacing and word choice
- Chat with an AI that feels like a supportive conversation partner
- Reflect, without any pressure
And because it’s all local, your privacy stays safe.
Prompting as Design: How We Shape AI to Feel Human
One big lesson from this project is that prompting isn’t just about asking questions—it’s a kind of design.
In regular apps, design means buttons, layouts, and workflows. But with AI, the prompt is part of the design—it’s the voice and tone of the experience.
That means writing prompts needs more than just technical skill. It needs:
Empathy: Knowing how the user might feel and using language that’s kind, respectful, and encouraging
Clarity: Being super clear about what you want the AI to do (and not do)
Cultural sensitivity: Making sure prompts respect language differences and user backgrounds
Lots of tweaking: Trying things out, learning from what the AI says, and adjusting the prompts until they sound natural and helpful
Some Prompt Examples from the Project
Rephrasings: Instead of telling the AI to “rewrite this sentence,” I ask for three short, easy-to-say alternatives that avoid tricky sounds. The prompt reminds the AI to be kind and accessible, not perfect or rigid.
Pacing advice: The AI gently suggests where to pause for breath, stretch vowels, or speak clearly. It’s practical and warm, not clinical.
Roleplay: The AI acts like a friendly conversation buddy, giving natural and encouraging replies that aren’t too long or formal.
Basically, prompts are the bridge between what people want and what the AI says. Thoughtful prompts make AI feel less like a cold machine and more like a caring helper.
Working with Local Tools and Close Loops
I built everything using Cursor, a conversational code editor, and Ollama running llama3 locally. I used FastApi for backend and React for frontend.
Why Cursor?
Cursor became a natural choice because it treats coding like a conversation. Instead of writing in isolation and switching constantly between code, terminal, and docs, I could ask, test, and refactor ideas in-line—almost like pair programming with an assistant who understood the context.
For a project rooted in prompting, where small shifts in language can change an entire outcome, this ability to iterate quickly and conversationally made all the difference.
Why LLaMA 3 via Ollama?
LLaMA 3 struck the right balance between quality and performance, especially when fine-tuned or guided through prompt engineering. Running it through Ollama offered a clean, developer-friendly way to manage and test models locally—crucial for keeping everything on-device. This wasn’t just about performance or cost.
It was about ownership: having a private, offline AI partner that respected the sensitivity of voice data and didn’t rely on cloud APIs or external tracking. It made testing feel safe and personal—closer to a local coach than a remote server.
Why FastAPI?
FastAPI offered just the right blend of modern Python tooling, speed, and clarity. Its support for async requests and automatic documentation meant I could move fast without sacrificing structure.
For a lightweight app that needed to listen to user input, query the model, and respond cleanly—all while remaining easy to test—FastAPI hit the sweet spot. It also made local deployment frictionless and kept the project small, portable, and maintainable.
This setup had great perks:
Privacy first: Nothing leaves your machine
Fast iteration: Try a prompt, tweak it, test it again
Conversational debugging: When something goes wrong, you get a back-and-forth instead of a frustrating error
Prompt and code shaping side by side: You see how prompt changes affect the app instantly
It made developing feel like a real conversation between me and the AI.
Designing for Emotional Safety
When working on a tool that touches something as personal as communication, emotional safety becomes as important as functionality. The goal wasn’t to evaluate or correct—it was to support without pressure.
That meant making deliberate choices in the user experience. The app avoids anything that might introduce anxiety or self-comparison:
No correctness scores
No tracking dashboards
No backend logging
No AI judgment
Just gentle suggestions, friendly roleplay, and a chance to reflect peacefully.
When Images Speak Softly
Design doesn’t stop at words. Visuals carry emotional weight too.
To reflect the same calm, inviting tone in the interface, I used ChatGPT’s image generation capabilities, crafting prompts that focused on emotional texture rather than glossy UI. The themes included:
“Engaging with Technology in Harmony”
“Speech Therapy Across Generations”
“Vibrant Microphones and Speech Pattern”
These AI-made images replaced generic stock photos and helped create a welcoming, inclusive vibe.
What I Learned Along the Way
This project deepened something I already sensed: prompting isn’t just technical—it’s emotional design. The way you ask shapes how users feel. A prompt can be cold and robotic, or kind and welcoming. And that tone becomes the voice of the experience itself.
Thoughtful prompts foster psychological safety.
AI-generated images, when guided with care, invite warmth and inclusion.
Local models protect privacy and create space to experiment without fear.
Together, these insights show AI as a partner, and not just a tool in supporting personal growth.
Final Thoughts
This project is one of many quiet experiments happening at the edges of the AI boom. Its biggest value is in the questions it raises and the perspective it offers.
As we build AI-powered tools, the most important things might not be features or apps, but how we build them: with thoughtful prompting, human-machine collaboration, and a real desire to create safe, caring spaces.
Even a small, local prototype made with attention and intention can shine a light on how technology might help us grow with patience, kindness, and humanity.
This content originally appeared on DEV Community and was authored by ujjavala