Resources AI Chatbot Plugin

Introduction

Beginners often struggle to take their first steps with Jenkins’ documentation and available resources. To address this challenge, this plugin integrates an AI-powered assistant directly into the Jenkins interface. It offers quick, intuitive support to users of all experience levels through a simple conversational UI.

The plugin is designed to reduce the learning curve for newcomers while also improving accessibility and productivity for experienced users.

This plugin was developed as part of a Google Summer of Code 2025 project.

Prerequisites

• Python: 3.11+
• Build tools: make, cmake (≥3.14), C/C++ compiler (gcc/clang/MSVC)
• Java: JDK 11+ and Maven 3.6+ (for plugin development)

Example system package installation

# Ubuntu/Debian/WSL
sudo apt install -y make cmake gcc g++ python3.11 python3.11-venv python3.11-dev

# macOS
brew install cmake python@3.11 && xcode-select --install

Getting Started

There are two ways to run the API locally, depending on your use case:

Option 1: Lite Mode (Recommended for Most Contributors)

Use this if you're working on the API, backend logic, data pipeline, or tests and don't need to test the actual chatbot responses.

make dev-lite

This will:

• Set up the Python environment automatically
• Install dependencies (skips the 4GB model download)
• Start the API server without loading the LLM

The API will be available at http://127.0.0.1:8000 within a few minutes.

Verify it's working:

curl -X POST http://127.0.0.1:8000/api/chatbot/sessions

What works: All API endpoints, session management, context search, data pipeline
What doesn't work: Actual chat completions (no model loaded)

Option 2: Full Mode (For Testing Chat Functionality)

Use this if you need to test the chatbot with real LLM responses or work on model-specific features.

First, complete the full setup in docs/setup.md. This includes installing llama-cpp-python and downloading the 4GB model.

Then run:

make api

The API will be available at http://127.0.0.1:8000.

What works: Everything, including actual chat completions with the local LLM

---

See docs/README.md for detailed explanations.

🎥 Setup Video Tutorial

The tutorial shows how to fork the repo, set up the backend, download the LLM model, run the frontend, and verify the chatbot works.

Troubleshooting

Model Downloads

• Symptom: The application appears "stuck" or frozen during the first run of the data pipeline or API.
• Cause: The system is downloading the embedding model (all-MiniLM-L6-v2, ~80MB) or initializing the LLM.
• Solution: This is normal behavior for the first run. Please wait for a few minutes. Ensure you have a stable internet connection.

Python Version Mismatches

• Symptom: SyntaxError or ModuleNotFoundError during setup or execution.
• Solution:
  • Ensure you are using Python 3.11+. Verify with python --version.
  • Ensure the virtual environment is activated:

    source chatbot-core/venv/bin/activate

Common Startup Errors

• Memory Limits: If the process is killed (e.g., OOM Killed), ensure your machine has sufficient RAM (at least 8GB recommended for full mode). Try running in Lite Mode (make dev-lite) first.
• Missing Dependencies: If you see import errors, re-run dependency installation:

  pip install -r chatbot-core/requirements.txt

• llama-cpp-python installation fails: Ensure build tools (gcc, cmake) are installed. See docs/setup.md for platform-specific instructions.

Verification Steps

To confirm your local setup is correct:

1. Virtual Environment: Ensure (venv) appears in your terminal prompt.
2. Lite Mode Check: Run make dev-lite. It should start without errors.
3. API Check: Run curl -X POST http://127.0.0.1:8000/api/chatbot/sessions. It should return a default session response.

For more details, see docs/setup.md.

Developer Documentation

Development-related documentation can be found in the docs/ directory.

Contributing

Refer to our contribution guidelines

LICENSE

Licensed under MIT, see LICENSE