Bee Agent Framework Starter

This starter template allows you to easily start working with the Bee Agent Framework in a second.

Key Features

• ✨ Safely execute an arbitrary Python Code via Bee Code Interpreter.
• ✨ Get complete visibility into agents’ decisions using our MLFlow integration thanks to Bee Observe.
• ✨ Fully fledged TypeScript project setup with linting and formatting.

Getting started

1. Clone the repository git clone git@github.com:i-am-bee/bee-agent-framework-starter or create your own repository from this one.
2. Install dependencies npm ci.
3. Fill missing values in .env.
4. Run the agent npm run start (it runs the ./src/agent.ts file).

Infrastructure

[!NOTE]

Docker distribution with support for compose is required, the following are supported:

• Docker
• Rancher - macOS users may want to use VZ instead of QEMU
• Podman - requires compose and rootful machine (if your current machine is rootless, please create a new one)

Code interpreter

Instructions

1. Start all services related to Code Interpreter npm run infra:start --profile=code_interpreter
2. Add CODE_INTERPRETER_URL=http://127.0.0.1:50051 to your .env (if .env does not exist, create one from .env.template).
3. Run the agent npm run start:code_interpreter (it runs the ./src/agent_code_interpreter.ts file)

Observability

Get full visibility of the agent's inner working via our observability stack.

• The MLFlow is used as UI for observability.
• The Bee Observe is the main Open-source observability service for Bee Agent Framework.
• The Bee Observe Connector is the observability connector for Bee Agent Framework

Configuration (ENV variables) can be found here.

Instructions

1. Start all services related to Observe npm run infra:start --profile=observe
2. Start the agent using the observe and MLFlow npm run start:observe (it runs the ./src/agent_observe.ts file).
3. Run the curl command that retrieves data from Bee Observe and passes them to the MLFlow instance.
4. Access MLFlow web application http://localhost:8080/#/experiments/
   • Credentials: (user: admin, password: password)