AgentUI

An Open Source AgentUI for your AgentOS

Kern provides a beautiful UI for interacting with your agents, completely open source, free to use and build on top of. It's a simple interface that allows you to chat with your agents, view their memory, knowledge, and more.

Note

The AgentOS only uses data in your database. No data is sent to Kern.

Built with Next.js and TypeScript, the Open Source Agent UI was developed in response to community requests for a self-hosted alternative following the success of AgentOS.

Get Started with Agent UI

To clone the Agent UI, run the following command in your terminal:

1npx create-agent-ui@latest

Enter y to create a new project, install dependencies, then run the agent-ui using:

1cd agent-ui && npm run dev

Open http://localhost:3000 to view the Agent UI, but remember to connect to your local agents.

Clone the repository manually▼

You can also clone the repository manually

1git clone https://github.com/kern-agi/agent-ui.git

And run the agent-ui using

1cd agent-ui && pnpm install && pnpm dev

Connect your AgentOS

The Agent UI needs to connect to a AgentOS server, which you can run locally or on any cloud provider.

Let's start with a local AgentOS server. Create a file agentos.py

1from kern.agent.agent import Agent
2from kern.models.openai import OpenAIChat
3from kern.os import AgentOS
4from kern.db.sqlite import SqliteDb
5from kern.tools.duckduckgo import DuckDuckGoTools
6from kern.tools.yfinance import YFinanceTools
7
8agent_storage: str = "tmp/agents.db"
9
10web_agent = Agent(
11    name="Web Agent",
12    model=OpenAIChat(id="gpt-4o"),
13    tools=[DuckDuckGoTools()],
14    instructions=["Always include sources"],
15    # Store the agent sessions in a sqlite database
16    db=SqliteDb(db_file=agent_storage),
17    # Adds the current date and time to the context
18    add_datetime_to_context=True,
19    # Adds the history of the conversation to the messages
20    add_history_to_context=True,
21    # Number of history responses to add to the messages
22    num_history_runs=5,
23    # Adds markdown formatting to the messages
24    markdown=True,
25)
26
27finance_agent = Agent(
28    name="Finance Agent",
29    model=OpenAIChat(id="gpt-4o"),
30    tools=[YFinanceTools()],
31    instructions=["Always use tables to display data"],
32    db=SqliteDb(db_file=agent_storage),
33    add_datetime_to_context=True,
34    add_history_to_context=True,
35    num_history_runs=5,
36    markdown=True,
37)
38
39agent_os = AgentOS(agents=[web_agent, finance_agent])
40app = agent_os.get_app()
41
42if __name__ == "__main__":
43    agent_os.serve("agentos:app", reload=True)

In another terminal, run the AgentOS server:

Setup your virtual environment

1python3 -m venv .venv
2source .venv/bin/activate

1python3 -m venv aienv
2aienv/scripts/activate

Install dependencies

1uv pip install -U openai ddgs yfinance sqlalchemy 'fastapi[standard]' kern-ai

1uv pip install -U openai ddgs yfinance sqlalchemy 'fastapi[standard]' kern-ai

Export your OpenAI key

1export OPENAI_API_KEY=sk-***

1setx OPENAI_API_KEY sk-***

Run the AgentOS

1python agentos.py

TipMake sure the module path in agent_os.serve() matches your filename (e.g., "agentos:app" for agentos.py).

View the AgentUI

Open http://localhost:3000 to view the Agent UI
Enter the localhost:7777 endpoint on the left sidebar and start chatting with your agents and teams!

Learn more

server

AgentOS Introduction

Learn about AgentOS

robot

Building Agents

Build your own agents