PersonaForge

• 

  structured chat and insight panel

• 

  default persona selection

• 

  create custom persona; flexibility, real use beyond default persona

INSPIRATION

Developers and product teams frequently build and ship ideas without early, targeted feedback. Existing validation methods are either slow, expensive, or rely on generic AI responses that don’t reflect real user perspectives. PersonaForge was created to simulate realistic, role-specific feedback so builders can validate ideas early and reduce wasted effort.

WHAT IT DOES

PersonaForge lets users talk to real people before they exist.

It enables users to:

• Chat with four default personas: Casual Friend, Technical Expert, Professional Mentor, and Creative Writer

• Create custom personas by defining tone, personality, expertise, and behavioral guidelines

• Maintain persona-specific memory for consistent responses across conversations

• View structured insights in a dedicated side panel, separate from chat

• Rapidly validate ideas without recruiting live testers

HOW WE BUILT IT

Frontend: React

Backend: Golang(Gin) Swagger for documentation, MongoDB as database, Render as free hosting service for the backend.

AI: Google Gemini API for the chat powered interface

Architecture: Persona-isolated sessions to prevent context leakage

UX Design: Split chat and insights panels to separate conversation from analysis

CHALLENGES WE RAN INTO

• Preventing memory and context leakage between multiple personas

• Designing prompts that produce realistic, non-generic feedback

• Managing response latency and rate limits with Gemini 3

• Presenting insights without overwhelming the user interface

ACCOMPLISHMENTS THAT WE'RE PROUD OF

• Fully functional multi-persona system with consistent memory

• Clear separation of conversational responses and structured insights

• Four distinct default personas with recognizable behavior patterns

• Stable integration of Gemini 3 in a real-world application context

WHAT WE LEARNED

• Persona realism depends heavily on prompt constraints and memory handling

• Separating insights from chat significantly improves usability

• AI feedback becomes more valuable when grounded in defined roles

• Clear MVP scoping is critical in time-constrained build

WHAT'S NEXT FOR PERSONAFORGE

• Persona marketplace for shared and reusable templates

• Collaborative testing for teams and startups

• Aggregated insight analytics across multiple sessions

• Improved mobile experience and performance optimization

Built With

• gemini-api
• golang
• javascript
• mongodb
• react
• render
• swagger
• vercel