Build a Cloud-based Ollama-powered Chat Interface with Full Authentication Integration

[syed-reza98]

Summary

Develop a fully integrated chat interface for the StormCom project that uses Ollama via a cloud endpoint as the backend LLM provider. The interface must leverage the existing authentication system, ensuring only authenticated users can access and use the chat functionality.

Objectives

• Create a modern, responsive chat UI within the Next.js 16 (App Router) framework, accessible only to authenticated users.
• Integrate with Ollama's cloud API (LLM backend) for real-time LLM-powered conversational responses.
• Ensure all messages and context are securely associated with authenticated sessions.
• Design for extensibility to support future multi-tenancy and fine-grained role-based access.

Requirements

1. UI/UX

• Add a new /chat protected route (App Router group) only accessible to logged-in users.
• Modern chat interface styled with shadcn-ui primitives; supports Markdown, code blocks, and loading states.
• Display user and assistant messages clearly, with session state persisted per user.
• Loading and error feedback for LLM response delays.

2. Authentication Integration

• Gate the entire chat route using the existing NextAuth middleware and session patterns.
• On the server: validate session tokens/JWT and associate chat histories with session.user.id.
• Support signed-in via email (current) and password (future); ensure session/tenant isolation.

3. Backend Integration (Ollama Cloud)

• Add secure backend API route /api/chat/ollama that:
  • Authenticating the user session on each request.
  • Accepts user message and sends it to configured Ollama cloud API.
  • Streams or returns LLM responses as appropriate.
• API keys for Ollama (and any secrets) must be set via .env.local and never exposed in frontend code.
• Add README documentation for required configuration, environment variables, and deployment steps.

4. Data Persistence

• Store chat history per user (PostgreSQL via Prisma 7):
  • Schema: User ID, message content (user/assistant), timestamps.
  • Model extensible to support organizations (multi-tenant) in future.

5. Security & Access

• Ensure only the authenticated user (or tenant org) may access their message history.
• Sanitize and validate all message inputs/outputs for safety (prevent prompt/HTML injection).

6. Dev Experience

• Include robust TypeScript typings throughout the chat, backend handler, and Prisma models.
• Ensure all new code passes npm run type-check and npm run lint (warnings OK, no errors).
• Add detailed documentation to the repository: usage instructions, config, and extending the chat.

Out of Scope

• No open anonymous usage; chat is for signed-in users only.
• No public endpoints to Ollama or chat API.
• No file uploads or image generation (text-only chat).
• Testing suite not required unless requested.

Acceptance Criteria

• Authenticated users can access /chat, send messages, and receive streaming responses from Ollama (cloud backend).
• All chat data persists securely in the database, scoped to user session.
• No sensitive credentials are leaked to client-side.
• Codebase passes build/type-check/lint workflows.
• Documentation updated for setup, environment config, and feature usage.

---

References:

• Ollama cloud API docs
• NextAuth.js Documentation
• Prisma Docs
• See /src/app/(auth) for authentication patterns.

---

Please ask in comments if further clarification is needed, or if there are specific product requirements for multi-tenancy and permissions.

[github-actions]

Automated review (GitHub Models):

A repo search found no evidence of a /chat route, backend API integration with Ollama, updated Prisma schema, or related documentation. The issue is likely not resolved yet.

Confidence: 0.7

Evidence:

• src/app/chat/page.tsx : No such file found; the protected /chat UI does not appear to exist.
• src/app/api/chat/ollama.ts : No backend API route for Ollama chat found in API directory.
• prisma/schema.prisma : No schema changes for user chat history (UserID, message, timestamps) per the requirements.
• README.md : No updated documentation for chat configuration, environment variables, or deployment steps.