[Feature]: Add `gitbun init` setup wizard that detects the project's framework, sets up `.gitbunrc`, and configures the optimal AI provider and model for the repository

[divyanshim27]

Summary

First-time Gitbun users must manually create a .gitbunrc or .smartcommitrc configuration file. The README provides examples but does not walk through what settings to choose for different project types. A gitbun init setup wizard would auto-detect the project type (Node.js, Python, Rust, etc.) and Ollama availability, then write a sensible default config — dramatically reducing onboarding friction for new GSSoC contributors and general users.

Problem

• New users running npx gitbun for the first time get the rule-based fallback if Ollama is not installed and no API key is configured — with no guidance on how to enable AI.
• The config format (.smartcommitrc vs .gitbunrc) is inconsistently named — README.md shows both, creating confusion about which one takes precedence.
• There is no way to discover available Ollama models from within Gitbun — users must separately run ollama list and then manually type the model name into their config.
• The VS Code extension (gitbun-vscode/) and the frontend playground (frontend/) are separate sub-projects — a user working in the VS Code extension context might need different configuration than a CLI user.

Proposed Solution

A gitbun init command that runs a guided setup:

// src/commands/init.ts
import inquirer from 'inquirer';
import { execSync } from 'child_process';
import { writeFileSync, existsSync } from 'fs';
import chalk from 'chalk';

async function detectOllamaModels(): Promise<string[]> {
  try {
    const output = execSync('ollama list --format json', { stdio: ['pipe', 'pipe', 'ignore'] }).toString();
    const models = JSON.parse(output);
    return models.map((m: { name: string }) => m.name);
  } catch {
    return [];
  }
}

function detectProjectType(): string {
  if (existsSync('Cargo.toml')) return 'rust';
  if (existsSync('pyproject.toml') || existsSync('requirements.txt')) return 'python';
  if (existsSync('go.mod')) return 'go';
  if (existsSync('package.json')) {
    const pkg = JSON.parse(readFileSync('package.json', 'utf-8'));
    if (pkg.dependencies?.next || pkg.devDependencies?.next) return 'nextjs';
    if (pkg.dependencies?.react || pkg.devDependencies?.react) return 'react';
    return 'nodejs';
  }
  return 'generic';
}

export async function runInitWizard() {
  console.log(chalk.cyan('\n🐇 Welcome to Gitbun! Let\'s set up your config.\n'));

  const projectType = detectProjectType();
  const ollamaModels = await detectOllamaModels();
  const hasOllama = ollamaModels.length > 0;

  console.log(chalk.gray(`Detected project type: ${projectType}`));
  console.log(chalk.gray(`Ollama: ${hasOllama ? `✓ (${ollamaModels.length} models available)` : '✗ not detected'}\n`));

  const answers = await inquirer.prompt([
    {
      type: 'list',
      name: 'provider',
      message: 'Which AI provider would you like to use?',
      choices: [
        ...(hasOllama ? [{ name: `Ollama (local, private)`, value: 'ollama' }] : []),
        { name: 'OpenAI (cloud, requires OPENAI_API_KEY)', value: 'openai' },
        { name: 'Anthropic Claude (cloud, requires ANTHROPIC_API_KEY)', value: 'anthropic' },
        { name: 'Rule-based only (no AI)', value: 'none' },
      ],
    },
    {
      type: 'list',
      name: 'model',
      message: 'Which Ollama model?',
      choices: ollamaModels,
      when: (a) => a.provider === 'ollama' && hasOllama,
    },
    {
      type: 'confirm',
      name: 'interactive',
      message: 'Always show preview before committing?',
      default: true,
    },
    {
      type: 'confirm',
      name: 'conventionalCommits',
      message: 'Enforce Conventional Commits format?',
      default: true,
    },
  ]);

  const config = {
    provider: answers.provider === 'none' ? undefined : answers.provider,
    model: answers.model,
    ai: answers.provider !== 'none',
    interactive: answers.interactive,
  };

  const configPath = '.gitbunrc';
  writeFileSync(configPath, JSON.stringify(config, null, 2));

  console.log(chalk.green(`\n✓ Config written to ${configPath}`));
  console.log(chalk.cyan('Run `gitbun` in any repo to generate your first commit message.\n'));
}

Additional Notes

• If .gitbunrc already exists, gitbun init will ask before overwriting it.
• The wizard auto-detects available Ollama models by calling ollama list — if Ollama is not installed, the Ollama option is hidden.
• The final config is always shown to the user for review before writing.
• gitbun init --yes (non-interactive mode) accepts all defaults for CI/scripted setups.
• I will add this to the README.md Quick Start section and update the CONTRIBUTING.md setup steps to recommend running gitbun init first.

Could you assign this issue to me?

Labels: enhancement, feature, dx, good first issue, GSSoC 2026

[divyanshim27]

Hey @nirvik34 ,
Please assign it to me under GSSoC 2026.
/claim /claim-assignment /assign