[Views] Initial ChatGPT-like chat view

[EtoboAgent]

Problem to solve

Based on PRD #1 (Knowforge), we need a first conversation interface to validate the product's core experience: asking questions and receiving answers in chat format.

As a workspace user, I want a simple and clear chat view to send questions and review message history, so I can interact with the knowledge base with minimal friction.

User experience goal

The user should be able to:

1. Open a chat screen.
2. View message history in chronological order.
3. Type a prompt in a bottom input area.
4. Send a message with a button or Enter key.
5. See a "thinking/responding" state while waiting for the reply.

Proposal

Create an initial frontend in @knowforge/views with a single ChatGPT-like view (visual MVP + basic flow):

• Two-zone layout:
  • Main area with message list.
  • Fixed bottom area with input and "Send" button.
• Different message bubbles for user and assistant.
• Auto-scroll to latest message.
• Local message state (no real backend in this first iteration).
• Mock assistant response with short delay to test UX.
• Structure ready to connect to a real chat endpoint in the next iteration.

Further details

Scope for this phase (initial frontend):

• Only one functional chat view.
• No authentication.
• No multi-tenant handling in the UI.
• No file upload or citations yet.
• No analytics.

Out of immediate scope:

• Conversation sidebar.
• Persistent user conversation history.
• Token-by-token streaming.
• Full RAG integration.

Permissions and Security

For this initial local issue, a full permissions model is not required. The view will be built locally using mock data.

Feature Usage Metrics

Minimum metrics to validate local UX:

• Time to first sent message.
• Number of messages sent per session.
• Empty-send attempt rate (input validation).

What does success look like, and how can we measure that?

Acceptance criteria:

• A chat route/screen exists and is accessible in the frontend.
• A user message can be sent and rendered in the list.
• A mock assistant response is rendered.
• There is visual feedback while waiting for response.
• The UI is usable on desktop (common laptop width).

Success definition (initial phase):

• Full chat flow (question -> response) manually validated with no console errors.

What is the competitive advantage or differentiation for this feature?

This validates the PRD's core experience from day one (knowledge chat) with a simple UI foundation that is fast to iterate and ready to evolve toward context-grounded answers with citations.

Links / references

• Main PRD: https://github.com/etobo-tech/knowforge/issues/1
• Scope: ChatGPT-like frontend MVP (first local iteration)