Gemini Golang Boilerplate

A lightweight starting point for building Go applications powered by the Gemini LLM. It ships with helpers for history management, response parsing, and a few prompt examples so you can focus on your idea.

Quick Start

1. Copy .env.example to .env and add your GEMINI_API_KEY (grab one from Google AI Studio).
2. Run go run . and type your prompt when asked.

The boilerplate includes three ready‑to‑use prompt handlers:

• BasicPersonalAssistant – converse with the model as a friendly assistant.
• PythonInterpreter – execute Python commands and receive the result as JSON.
• TalkFile – ask questions about the contents of a file wrapped between [content_filename=...] tags.

Choose one in main.go or create your own.

Creating a Custom Prompt

Add a new file in prompts/ that formats the user input into the instructions you want to send to Gemini. A minimal handler looks like this:

// prompts/my_prompt.go
package prompts

import (
    "app/gemini"
    "fmt"
)

var TEMPLATE = `Context: %s
User Prompt: %s`

func MyPrompt(userPrompt string) string {
    return fmt.Sprintf(TEMPLATE, gemini.BuildHistory(), userPrompt)
}

Update main.go to call prompts.MyPrompt and set jsonMode depending on your output format.

Processing Responses

processResponse can optionally parse JSON and stores the conversation history for you. The history length defaults to 100 messages but can be changed in gemini/history.go.

Why Use This Boilerplate?

• Quickly experiment with Gemini using concise Go code.
• Extend or replace the existing prompts to fit your own workflows.
• Built‑in helpers remove repetitive boilerplate so you can test ideas faster.

Start hacking and let Gemini power your next project!