Simple LLM Chatbot App
This is a basic conversational chatbot built using Zedflows. It uses a 2-node workflow to power a personalized assistant that remembers previous messages per user.
What It Does
- Takes user input via a text field
- Sends the message to an LLM
- Maintains a conversation using memory (per user)
- Replies back with a contextual response
Workflow Setup
The app consists of a single workflow with 2 main nodes:
1. Trigger Node: Text Input
- Presents a chat input field to the user
- Accepts their message (e.g. "Tell me a joke")
- Passes this text to the next node
- Each input is scoped to the user (via session or ID)
2. LLM Node: Chat with Memory
- Connected to the trigger
- Uses a Memory Tool to keep conversation history per user
- Memory scope is user-specific, so each session is personalized
- Supports multiple turns in the same context
How to Build It in Zedflows
- Create a new Workflow
- Add a Trigger Node
- Choose type:
Text Input
- Customize placeholder: "Ask me anything..."
- Add an LLM Node
- Set prompt: "You are a helpful assistant. Answer conversationally."
- Enable
Memory Tool
- Connect the Trigger → LLM
- Optionally create an App
🌐 Result
Users visit your app, type a message, and get a smart response. The assistant remembers what was said earlier in the chat, so it can hold natural conversations.
This is the perfect base to:
- Expand with APIs or vector search
- Turn into a customer support bot