A streaming chat web UI for an agentic RAG assistant — renders responses token-by-token over SSE, backed by the agentic-rag server.
Start server on http://localhost:3000
pnpm dev/ - landing page
/chat/[id] - streaming chat UI
agentic-chat project is specifically built to connect with agentic-rag repo's API server.
To establish the connection locally, load qwen2.5:7b model via ollama on your laptop first.
ollama run qwen2.5:7bYou should see something like this in terminal:
➜ ~ ollama run qwen2.5:7b
>>> Send a message (/? for help)
Next, clone agentic-rag repo, and launch the following servers locally.
In agentic-rag/, run
# mcp server
uv run -m mcp_server.server
# fastapi streaming chat server
fastapi dev --entrypoint api.server:app --port 8001