devcoach
Local MCP server that adds one short technical lesson after each agent task
- Pricing
- Open source
- Platforms
- CLI, macOS, Linux
- Runs
- Locally
- Version
- 2.7.1
- Listing updated
- 1 day ago
About devcoach
devcoach is an open-source MCP server for education and exams. It runs locally on your machine and is started by your AI client.
devcoach is built on the idea that people ship more with AI agents while understanding less. After your agent finishes a task, it appends a short lesson on something you probably do not know yet, such as when to use Promise.allSettled instead of Promise.all, labelled by topic and level.
The first connection scans local Claude Code history, using file and activity metadata, to build a knowledge map and a coaching notebook with no questions asked. An opt-in deeper tier also reads local session transcripts. Under each lesson you reply y if you knew it, u if it was new and clear, or n if it was too dense, and confidence on that topic adjusts. By default there are at most two lessons a day, at least four hours apart, and none in plan mode.
A local dashboard opened with devcoach ui shows topics, lesson history and pacing. It is free, open source and local, installable via npm, Homebrew or a Claude Desktop extension.
Install
Run in a terminal
claude mcp add devcoach -- npx -y devcoachAdd to claude_desktop_config.json
{
"mcpServers": {
"devcoach": {
"command": "npx",
"args": [
"-y",
"devcoach"
]
}
}
}Add to ~/.cursor/mcp.json
{
"mcpServers": {
"devcoach": {
"command": "npx",
"args": [
"-y",
"devcoach"
]
}
}
}Add to .vscode/mcp.json
{
"servers": {
"devcoach": {
"type": "stdio",
"command": "npx",
"args": [
"-y",
"devcoach"
]
}
}
}Add to ~/.codeium/windsurf/mcp_config.json
{
"mcpServers": {
"devcoach": {
"command": "npx",
"args": [
"-y",
"devcoach"
]
}
}
}Add to your client's MCP settings (most use this shape)
{
"mcpServers": {
"devcoach": {
"command": "npx",
"args": [
"-y",
"devcoach"
]
}
}
}What it does
- In-context lessons
- Appends a short lesson after a completed task.
- Knowledge map
- Builds a topic and confidence map from local history.
- One-key feedback
- Reply y, u or n to tune what comes next.
- Quiet pacing
- Defaults to two lessons a day, hours apart.
- Local dashboard
- Browse lessons and adjust confidence.
- Multiple agents
- Works with Claude, Gemini CLI, Codex CLI and Cursor.
Good for
- Learn from the code your agent just wrote
- Track which topics you already know
- Limit coaching to a few lessons per day
- Review lesson history in a local dashboard
Pricing
Open source
Free, open source, runs locally
Prices change. Check ultimaphoenix.github.io before you buy.
devcoach FAQ
Does it send data anywhere?
The site describes it as local and free.
How often are lessons shown?
At most 2 a day by default, at least 4 hours apart.
What does onboarding read?
Local Claude Code history metadata; a deeper tier also reads session transcripts.
Which clients work?
Claude Code, Claude Desktop, Gemini CLI, Codex CLI, Cursor and others.
Is devcoach free?
Yes. devcoach is open source, so you can use and self-host it for free.
How do I add devcoach to Claude Code?
Run this command in your terminal: claude mcp add devcoach -- npx -y devcoach
What are the best alternatives to devcoach?
The closest alternatives to devcoach are REA, Codebase Memory, GitHub, Serena MCP: the IDE for your agent and XERJ.
Alternatives to devcoach

REA
MCP serverCLI and MCP server that lets coding agents reverse engineer apps and binaries
Runs locally- Open source
- 464K GitHub stars
- Developer tools

Codebase Memory
MCP serverLocal MCP server that builds a code knowledge graph for AI coding agents
Runs locally- Open source
- 48.7K GitHub stars
- Knowledge and memory

GitHub
MCP serverOfficial MCP server connecting AI assistants to GitHub repos, issues and workflows
Not checked yet- Remote
- Open source
- 34.5K GitHub stars
- Developer tools

Serena MCP: the IDE for your agent
MCP serverMCP toolkit giving coding agents symbol-level code retrieval and editing
Runs locally- Open source
- 31.5K GitHub stars
- Developer tools

XERJ
MCP serverLocal search engine for AI agents with cited answers over your code and docs
Runs locally- Open source
- 19.2K GitHub stars
- Knowledge and memory

Ouroboros
MCP serverSpec-first layer that pins requirements and verifies results of AI coding agent work
Runs locally- Open source
- 6.4K GitHub stars
- Developer tools