About changelogai
changelogai is a CLI for generating release notes from Git history and suggesting semantic version bumps. It parses conventional commits, with optional Claude rewriting and an MCP server for agent and IDE workflows.
Key features
- 01Generate changelogs from Git history
- 02Group commits by category
- 03Rewrite notes with Claude
- 04Export Markdown, JSON, or plain text
- 05Suggest semantic version bumps
Use cases
- 01Draft release notes from the latest tag
- 02Prepare user-facing notes from technical commits
- 03Choose a major, minor, or patch release
MCP support
Its MCP server provides tools to generate changelogs in Markdown, JSON, or plain text, and to suggest a semantic version bump.
Configuration
Local server config is the same for every client. Copy it into your client’s MCP config file.
Tools
2Saved tool list · updated 2w ago
| Parameter | Type | Required | Description |
|---|---|---|---|
| to | string | No | End ref. Default: HEAD |
| from | string | No | Start ref (tag/commit). Default: latest tag |
| repo | string | No | Path to git repository |
| group | boolean | No | Group commits by type |
| format | enum | No | markdown | json | plain |
| maxCommits | number | No | Max commits to process |
Details
From the project READMEchangelogai
Generate professional changelogs from your git history — instantly.
Works out of the box with conventional commits. Optional AI mode uses Claude to rewrite technical commit messages into user-facing release notes.
Install
npm install -g @ura-dev/changelogai
Quick Start
# Run without installing
npx @ura-dev/changelogai
# Generate changelog since last tag
changelogai
# Grouped by category
changelogai --group
# AI-enhanced (requires ANTHROPIC_API_KEY)
changelogai --ai --group
# Write to file
changelogai --group --output CHANGELOG.md
# Prepend new release to existing changelog
changelogai --group --output CHANGELOG.md --prepend
# JSON output for CI/CD
changelogai --format json
# Suggest semver bump
changelogai --version-bump
Features
- Zero config — auto-detects latest tag, parses conventional commits
- Smart grouping — categorizes commits into Features, Bug Fixes, Performance, etc.
- AI mode — rewrites commit messages into clear, user-facing descriptions using Claude
- Multiple formats — Markdown, JSON, plain text
- Version bump suggestion — analyzes changes and suggests major/minor/patch
- Prepend mode — add new releases to the top of existing CHANGELOG.md
- No dependencies — pure Node.js, works everywhere
Options
| Flag | Description | Default |
|---|---|---|
--from <ref> | Start ref (tag/commit) | Latest tag |
--to <ref> | End ref | HEAD |
--output <file> | Write to file | stdout |
--format <fmt> | markdown, json, plain | markdown |
--group | Group by commit type | false |
--ai | AI-enhanced descriptions | false |
--model <model> | Claude model for AI mode | claude-haiku-4-5-20251001 |
--prepend | Prepend to existing file | false |
--version-bump | Suggest semver bump | false |
--repo <path> | Git repo path | . |
--max-commits <n> | Max commits to process | 500 |
MCP Server
changelogai includes a Model Context Protocol server for AI agents and IDE integrations (Claude Desktop, Cursor, VS Code, etc.).
{
"mcpServers": {
"changelogai": {
"command": "npx",
"args": ["-y", "-p", "@ura-dev/changelogai", "changelogai-mcp"]
}
}
}
Or if installed globally:
{
"mcpServers": {
"changelogai": {
"command": "changelogai-mcp"
}
}
}
Available Tools
| Tool | Description |
|---|---|
changelogai_generate | Generate a changelog from git history. Returns markdown, JSON, or plain text. |
changelogai_version_bump | Analyze git commits and suggest next semver bump (major/minor/patch). |
AI Mode
Set your Anthropic API key:
export ANTHROPIC_API_KEY=sk-ant-...
Then run with --ai:
changelogai --ai --group --output CHANGELOG.md
AI mode rewrites technical commit messages into clear, user-facing release notes. Uses Claude Haiku by default (fast and cheap — ~$0.001 per changelog).
License
MIT
Comments
New to MCP?
From what the protocol is to installing your first server, these answers cover most first questions.
An open protocol that defines how AI applications connect to external tools and data in a uniform way. Clients talk to the model; MCP servers provide tools, resources and prompt templates. They exchange standard messages, so one server can be reused by many clients.