Inspiration
Most AI coaching and productivity apps focus heavily on action items, plans, and rigid frameworks. That approach often creates pressure instead of clarity. Momentum Coaching AI was inspired by the belief that progress starts with conversation. People need a calm space to think, reflect, and talk things through before being told what to do.
Another key inspiration was the idea that coaching is personal. Different people respond to different tones, personalities, and styles. Instead of offering one generic assistant, Momentum Coaching AI allows users to create or generate coaches that actually fit how they think and communicate.
What it does
Momentum Coaching AI is a conversation first AI coaching app that lets users create personalized AI coaches and chat with them over time.
Users sign up and receive free tokens to explore the experience. They can generate multiple AI coaches based on their interests and goals or create one manually. Each coach has its own personality, style, category, and voice.
Once a coach is selected, users can have ongoing conversations with it. The coach responds based on its defined system prompt, remembers conversation context, and supports users without pushing rigid action plans or overwhelming advice.
How we built it
The app is built using a feature based architecture to keep the system modular and scalable.
The AI experience is powered by layered prompting. A global base system prompt defines app wide behavior and safety. Each coach has its own system prompt that defines personality and style. Dynamic runtime context such as user name, goals, and interests is injected into each chat request.
Firebase is used for authentication, database storage, image storage, and real time features. RevenueCat handles subscriptions and token based usage, supporting both recurring plans and one time token purchases.
Chat history is sent with each request to maintain conversational continuity, with safeguards to manage token usage.
Challenges we ran into
Designing prompts that felt human without becoming repetitive or overly directive was one of the biggest challenges. Separating global rules from coach specific behavior required careful iteration.
Another challenge was allowing users to freely create coaches while maintaining consistent quality and safety across conversations.
Balancing token usage, subscription pricing, and user experience without adding friction was also a key challenge.
Accomplishments that we’re proud of
We built a system where users can generate multiple unique coaches and choose the one that resonates with them.
The conversation experience feels natural, calm, and consistent across sessions.
The architecture allows new coaches, features, and monetization options to be added without reworking the core system.
Most importantly, Momentum Coaching AI avoids the common pitfall of AI coaching tools that feel pushy or overwhelming.
What we learned
Small changes in prompt design have a huge impact on how human an AI feels.
Users value tone, pacing, and feeling understood more than perfectly structured advice.
AI products are not just about intelligence, but about intentional constraints and thoughtful experience design.
Conversation design is as important as model choice.
What’s next for Momentum Coaching AI
Future versions of Momentum Coaching AI will gently bridge insight and action while preserving the conversation first experience.
We plan to introduce optional, user approved action suggestions generated from conversations, such as starting a focus timer, creating a lightweight task, or saving a reflection. These features will never be forced and will only appear when relevant.
We also plan to add deeper long term memory so coaches can better understand users over time.
Additional plans include more coach categories, better discovery, and more customization controls.
The long term vision is to make Momentum Coaching AI a trusted space for reflection, clarity, and steady momentum.
Log in or sign up for Devpost to join the conversation.