Zeltra AI
An AI chatbot for entrepreneurs and people thinking about starting a company.
- Status
- Live
- Role
- Full Stack Developer
- Team
- Solo
- Timeline
- 3 days
The problem
Founders start with vague ideas that need refining and validating.
Who it is for
Founders and entrepreneurs.
What I built
An AI chatbot that turns vague ideas into clear problem and solution statements, analyzes market niches, gives validation signals, and suggests MVP scope. Built with the Next.js App Router, LangChain.js with Gemini and Llama, Prisma and PostgreSQL, and NextAuth.
Built with
- Frontend
- Next.js
- TypeScript
- API & Auth
- Next.js App Router
- NextAuth
- Data
- PostgreSQL
- Prisma
- AI
- LangChain.js
- Gemini
- Llama
In detail
- Build timeline
- 3 days
- API endpoints
- 7
- LLM models
- 2
Note
Scope
Zeltra is still evolving and focuses on clarity and decision-making. It does not cover execution or full business planning.
Key features
Idea refinement
Turns vague startup ideas into clear problem and solution statements, so founders have a concrete starting point.
Market niche analysis
Identifies target users, pain points, and underserved opportunities around the idea.
Startup validation signals
Evaluates feasibility using demand, competition, and differentiation signals.
Prompt-based ideation engine
Generates structured prompts for deeper exploration of a concept.
Prototype direction generator
Suggests MVP scope, core features, and technical direction to move from idea to build.
AI-guided iteration
Improves ideas through repeated questioning rather than a single answer.
API overview
Zeltra AI's API is built on the Next.js App Router. Its endpoints cover the chat lifecycle, sessions, AI analysis, and authentication, and all routes live under src/app/api/*.
| Endpoint | Method | Purpose |
|---|---|---|
/api/new-chat |
POST | Creates a new chat session |
/api/get-messages |
GET | Fetches messages for an active chat session |
/api/delete-chat/{chatId} |
DELETE | Deletes a specific chat session by its unique ID |
/api/get-sessions |
GET | Retrieves all chat sessions associated with the authenticated user |
/api/analyze |
POST | Runs AI-powered analysis on user input: idea refinement, validation, structured reasoning |
/api/auto-initiate |
POST | Automatically initializes a chat session, used for first-time users and quick-start flows |
/api/auth/[...nextauth] |
GET, POST | Handles authentication with NextAuth: providers, sessions, and callbacks |
Route structure reference
api/
├── analyze/
├── auth/[...nextauth]/
├── auto-initiate/
├── delete-chat/[chatId]/
├── get-messages/
├── get-sessions/
└── new-chat/Challenges
- AI memory management
- Multi-model support
- Langchain integration
What I learned
- Chatbot Architecture
- Database Operations
Have a role in mind?
Send me a message on X with the role and a link to the posting. My resume has the rest.