Back to Apps

CodeBrain
by feelingai-team
LSP-powered codebase intelligence for AI agents, featuring structural search, validation, and monorepo support.
0 stars
Works in:claude
Exposes:ToolsResources
What it does
CodeBrain acts as a high-performance cognitive layer for AI agents working with code. It connects LLMs to a codebase using the Language Server Protocol (LSP), allowing agents to understand project structures, symbols, and dependencies with precision far beyond simple text search.
Tools
validate: Provides compiler-grade diagnostics to catch type errors and missing imports instantly.outline: Maps the file hierarchy and symbol tree of a codebase.search: Performs structural search by syntactic role instead of string matching.explore_symbol: Surfaces callers and usages to understand impact before refactoring.check_impact: Analyzes downstream breakage caused by potential code changes.debug_trace: Enriches stack traces with code intelligence to accelerate root cause analysis.rename_symbol: Safely renames symbols across a project.list_subprojects: Automatically discovers sub-project boundaries in monorepos.check_health: Reports the current capability and degradation state of the server.
Installation
Install the package via pip:
pip install "codebrain[all] @ git+https://github.com/feelingai-team/CodeBrain.git"
Configure in claude_desktop_config.json:
{
"mcpServers": {
"codebrain": {
"command": "codebrain-mcp",
"args": []
}
}
}
Supported hosts
- claude
- opencode
Quick install
pip install "codebrain[all] @ git+https://github.com/feelingai-team/CodeBrain.git"Information
- Pricing
- free
- Published
- 7/30/2026
- stars
- 0
Categories
Choose your AI client and follow the steps below.
Codex
Add `codebrain-mcp` to MCP config.Claude Desktop
{"mcpServers": {"codebrain": {"command": "codebrain-mcp", "args": []}}}






