CKB — Code Knowledge Backend
Know your code. Change it safely. Ship with confidence.
CKB transforms your codebase into a queryable knowledge base. Ask questions, understand impact, find owners, detect dead code—all through CLI, API, or AI assistants.
Think of it as a senior engineer who knows every line of code, every decision, and every owner—available 24/7 to answer your questions.
Instant Answers to Hard Questions
What You Can Do
🔍 Understand — Semantic search, call graphs, usage tracing, architecture maps
⚡ Analyze — Impact analysis, risk scoring, hotspot detection, coupling analysis
🛡️ Protect — Affected test detection, breaking change warnings, PR risk assessment
🔐 Secure — Secret detection, credential scanning, security-sensitive code identification
👥 Collaborate — Ownership lookup, reviewer suggestions, architectural decisions (ADRs)
📊 Improve — Dead code detection, tech debt tracking, documentation coverage
🚀 Compound Operations — Single-call tools (explore, understand, prepareChange) reduce AI tool calls by 60-70%
🔗 Integrate — CLI, HTTP API, MCP for AI tools, CI/CD pipelines, custom scripts
Try It Now
# See what's risky in your codebase
ckb hotspots --format=human
# Check what changed and what might break
ckb diff-summary --format=human
# Scan for exposed secrets
ckb audit --format=human
# Check architecture at a glance
ckb arch --format=human
# Check system status
ckb status
Works Everywhere
83% token reduction with smart presets—load only the tools you need.
# One command to connect to Claude Code
ckb setup
Building your own tools? Use CKB as a backend via CLI, HTTP API, or MCP. See the Integration Guide for examples in Node.js, Python, Go, and shell scripts.
Learn More
Quick Start
Option 1: npm (Recommended)
# Install globally
npm install -g @tastehub/ckb
# Or run directly with npx (no install needed)
npx @tastehub/ckb init
Option 2: Homebrew (macOS/Linux)
brew install simplyliz/ckb/ckb
Use the fully-qualified name. Homebrew 6.0 requires third-party taps to be trusted before it loads them; installing this way grants that trust for this one formula. Tapping first and installing by short name fails with
Refusing to load formula ... from untrusted tap.
Option 3: Build from Source
git clone https://github.com/SimplyLiz/CodeMCP.git
cd CodeMCP
go build -o ckb ./cmd/ckb
Setup
# 1. Initialize in your project
cd /path/to/your/project
ckb init # or: npx @tastehub/ckb init
# 2. Generate SCIP index (optional but recommended)
ckb index # auto-detects language and runs appropriate indexer
# 3. Connect to Claude Code
ckb setup # creates .mcp.json automatically
# Or manually:
claude mcp add --transport stdio ckb -- npx @tastehub/ckb mcp
Token efficiency shown at startup:
CKB MCP Server v9.3.1
Active tools: 25 / 110 (22%)
Estimated context: ~4k tokens
Preset: core
Now Claude can answer questions like:
- "What calls the HandleRequest function?"
- "How is ProcessPayment reached from the API?"
- "What's the blast radius if I change UserService?"
- "Who owns the internal/api module?"
- "Is this legacy code still used?"
Why CKB?
Three Ways to Use It
How Indexing Works
CKB uses SCIP indexes to understand your code. Think of it like a database that knows where every function is defined, who calls it, and how everything connects.
The Basics
# 1. Generate an index (auto-detects language)
ckb index
# 2. Check if your index is fresh
ckb status
Without an index, CKB still works using tree-sitter parsing (basic mode), but with an index you get:
- Cross-file references ("who calls this function?")
- Precise impact analysis
- Call graph navigation
Language Support
Not all languages are equal. CKB classifies languages into quality tiers based on indexer maturity:
Key limitations:
- Incremental indexing is Go-only. Other languages require full reindex.
- TypeScript monorepos may need
--infer-tsconfigflag - C/C++ requires
compile_commands.json - Python works best with activated virtual environment
Run ckb doctor --tier standard to check if your language tools are properly installed.
See Language Support for indexer installation, known issues, and the full feature matrix.
Keeping Your Index Fresh
Your index becomes stale when you make commits. CKB offers several ways to stay current:
Quick start for AI sessions:
ckb mcp --watch # Auto-reindexes every 30s when stale
Check staleness:
ckb status
# Shows: "5 commits behind HEAD" or "Up to date"
For Go projects, CKB uses incremental indexing—only changed files are processed, making updates fast.
See the Index Management Guide for complete documentation.
Features
📖 Full Features Guide — Detailed documentation with examples
📋 Changelog — Version history
PR Review
CKB review runs 21 quality checks in 5 seconds — secrets, breaking changes, dead code, complexity, test gaps, bug patterns, and more. Zero tokens, zero API calls.
When your AI assistant (Claude Code, Cursor, Windsurf) reviews a PR, it calls CKB first and gets structured analysis in ~1k tokens. Then it only reads the files that matter — saving 50-80% of tokens on large PRs.
ckb review --base=main # Human-readable review
ckb review --base=main --ci # CI mode (exit codes)
ckb review --base=main --post=123 # Post as PR comment
Works in CI without any LLM:
- run: npx @tastehub/ckb review --base=main --ci --format=sarif > review.sarif
📖 How it helps AI review · Benchmarks · CI Integration · Quickstart
CLI
ckb status # System health (with remediation suggestions)
ckb search Handler # Find symbols
ckb diff-summary # Analyze what changed
ckb hotspots # Risky areas
ckb arch # Architecture overview
ckb ownership # File/path ownership
ckb mcp # Start MCP server
Compound Operations (via MCP):
# These tools combine multiple queries into single calls
explore # Area exploration: symbols, dependencies, hotspots
understand # Symbol deep-dive: refs, callers, explanation
prepareChange # Pre-change analysis: impact, tests, risk
batchGet # Fetch up to 50 symbols at once
batchSearch # Run up to 10 searches at once
📖 User Guide — All CLI commands and options
HTTP API
# Start the HTTP server
ckb serve --port 8080
# Example calls
curl http://localhost:8080/health
curl http://localhost:8080/status
curl "http://localhost:8080/search?q=NewServer"
curl http://localhost:8080/architecture
curl "http://localhost:8080/ownership?path=internal/api"
curl http://localhost:8080/hotspots
# Index Server Mode (v7.3) - serve indexes to remote clients
ckb serve --port 8080 --index-server --index-config config.toml
# Index server endpoints
curl http://localhost:8080/index/repos
curl http://localhost:8080/index/repos/company%2Fcore-lib/meta
curl "http://localhost:8080/index/repos/company%2Fcore-lib/symbols?limit=100"
curl "http://localhost:8080/index/repos/company%2Fcore-lib/search/symbols?q=Handler"
# Upload endpoints (with compression + auth)
curl -X POST http://localhost:8080/index/repos \
-H "Content-Type: application/json" \
-H "Authorization: Bearer ckb_xxx" \
-d '{"id":"my-org/my-repo","name":"My Repo"}'
gzip -c index.scip | curl -X POST http://localhost:8080/index/repos/my-org%2Fmy-repo/upload \
-H "Content-Encoding: gzip" \
-H "Authorization: Bearer ckb_xxx" \
--data-binary @-
# Token management (index server admin)
ckb token create --name "ci-upload" --scope upload # Create API key
ckb token list # List all tokens
ckb token revoke ckb_xxx # Revoke a token
ckb token rotate ckb_xxx # Rotate (new secret, same ID)
MCP Integration
CKB works with any MCP-compatible AI coding tool.
# Auto-configure for current project
npx @tastehub/ckb setup
# Or add globally for all projects
npx @tastehub/ckb setup --global
Or manually add to .mcp.json:
{
"mcpServers": {
"ckb": {
"command": "npx",
"args": ["@tastehub/ckb", "mcp"]
}
}
}
Add to ~/.cursor/mcp.json (global) or .cursor/mcp.json (project):
{
"mcpServers": {
"ckb": {
"command": "npx",
"args": ["@tastehub/ckb", "mcp"]
}
}
}
Add to ~/.codeium/windsurf/mcp_config.json:
{
"mcpServers": {
"ckb": {
"command": "npx",
"args": ["@tastehub/ckb", "mcp"]
}
}
}
Add to your VS Code settings.json:
{
"mcp": {
"servers": {
"ckb": {
"type": "stdio",
"command": "npx",
"args": ["@tastehub/ckb", "mcp"]
}
}
}
}
Add to opencode.json in project root:
{
"mcp": {
"ckb": {
"type": "local",
"command": ["npx", "@tastehub/ckb", "mcp"],
"enabled": true
}
}
}
# Auto-configure for current project
npx @tastehub/ckb setup --tool=grok
# Or add globally
npx @tastehub/ckb setup --tool=grok --global
Or manually add to .grok/settings.json (project) or ~/.grok/user-settings.json (global):
{
"mcpServers": {
"ckb": {
"name": "ckb",
"transport": "stdio",
"command": "npx",
"args": ["@tastehub/ckb", "mcp"]
}
}
}
Claude Desktop doesn't have a project context, so you must specify the repository path.
Automatic setup (recommended):
cd /path/to/your/repo
ckb setup --tool=claude-desktop
Manual configuration — add to ~/Library/Application Support/Claude/claude_desktop_config.json (macOS) or %APPDATA%\Claude\claude_desktop_config.json (Windows):
{
"mcpServers": {
"ckb": {
"command": "npx",
"args": ["-y", "@tastehub/ckb", "mcp"],
"env": {
"CKB_REPO": "/path/to/your/repo"
}
}
}
}
The CKB_REPO environment variable tells CKB which repository to analyze. Claude Desktop can only work with one repository at a time.
Use cmd /c wrapper in any config above:
{
"mcpServers": {
"ckb": {
"command": "cmd",
"args": ["/c", "npx", "@tastehub/ckb", "mcp"]
}
}
}
CKB exposes 110+ tools, but most sessions only need a subset. Use presets to reduce token overhead by up to 77%:
# List all available presets with tool counts and token estimates
ckb mcp --list-presets
# Default: core preset (25 essential tools)
ckb mcp
# Workflow-specific presets
ckb mcp --preset=core # 25 tools - search, explain, impact (default)
ckb mcp --preset=review # 42 tools - core + diff, ownership, PR review
ckb mcp --preset=refactor # 42 tools - core + coupling, dead code
ckb mcp --preset=federation # 46 tools - core + cross-repo
ckb mcp --preset=docs # 34 tools - core + doc-symbol linking
ckb mcp --preset=ops # 43 tools - core + jobs, webhooks, metrics
ckb mcp --preset=full # 110 tools - complete feature set
In MCP config:
{
"mcpServers": {
"ckb": {
"command": "npx",
"args": ["@tastehub/ckb", "mcp", "--preset=review"]
}
}
}
The AI can dynamically expand the toolset mid-session using the expandToolset tool.
Under the Hood
CKB orchestrates multiple code intelligence backends, using whatever is available and stacking their strengths — more backends running means richer answers:
- Cartographer — Fast structural analysis (architecture, dependencies, blast-radius, coupling, directory rollup) that needs no index and works the moment CKB is installed. Compiled into the binary, so it ships in the npm and Homebrew builds by default.
- SCIP — Precise, pre-indexed symbol data for exact definitions, references, and call graphs (run
ckb index) - LSP — Real-time language server queries
- Git — Blame, history, churn analysis, ownership
- LIP — Semantic embedding daemon for nearest-neighbour search and re-ranking. Runs as a separate, optional daemon — install and start it alongside CKB.
Results are merged intelligently and compressed for LLM context limits. Every backend except Git is optional; CKB degrades gracefully to whatever is present.
LIP enhances semantic search, PR novelty detection, test discovery, file boundary analysis, and architecture coupling signals. When LIP is running alongside CKB, search quality improves significantly — especially for natural-language queries that don't match symbol names literally. See Hybrid Retrieval for details, or the LIP documentation.
Persistent knowledge survives across sessions:
- Module Registry — Boundaries, responsibilities, tags
- Ownership Registry — CODEOWNERS + git-blame with time decay
- Hotspot Tracker — Historical snapshots with trend analysis
- Decision Log — ADRs with full-text search
Who Should Use CKB?
- Developers using AI assistants — Give your AI tools superpowers
- Teams with large codebases — Navigate complexity efficiently
- Anyone doing refactoring — Understand impact before changing
- Code reviewers — See the full picture of changes
- Tech leads — Track architectural health over time
Limitations (Honest Take)
CKB excels at:
- Static code navigation—finding definitions, references, call graphs
- Impact analysis for safe refactoring
- Ownership lookup (CODEOWNERS + git blame)
- Architecture and module understanding
CKB won't help with:
- Dynamic dispatch / runtime behavior (use debugger)
- Generated code that isn't indexed
- Code generation, linting, or formatting
- Cross-repo calls (use federation for this)
CKB is static analysis, not magic. Always verify critical decisions by reading the actual code.
📖 Practical Limits — Full guide on accuracy, blind spots, and when to trust results
Documentation
See the Full Documentation Wiki for:
- Quick Start — Step-by-step installation
- Prompt Cookbook — Real prompts for real problems
- Language Support — SCIP indexers and support tiers
- Practical Limits — Accuracy notes, blind spots
- User Guide — CLI commands and best practices
- Index Management — How indexing works, auto-refresh methods
- Incremental Indexing — Fast index updates for Go projects
- Doc-Symbol Linking — Symbol detection in docs, staleness checking
- Authentication — API tokens, scopes, rate limiting
- MCP Integration — Claude Code setup, 110+ tools
- API Reference — HTTP API documentation
- Daemon Mode — Always-on service with scheduler, webhooks
- Configuration — All options including MODULES.toml
- Architecture — System design and components
- Security — Secret detection, credential scanning
- Telemetry — Runtime observability, dead code detection
- Federation — Cross-repository queries
- CI/CD Integration — GitHub Actions, PR analysis
Requirements
Using npm (recommended):
- Node.js 16+
- Git
Building from source:
- Go 1.21+
- Git
Optional (for enhanced analysis):
- SCIP indexer for your language (scip-go, scip-typescript, etc.) — run
ckb indexto auto-install - LIP v2.3+ semantic embedding daemon (requires Rust/cargo) — strongly recommended for best search quality (docs, crates.io):
cargo install lip-cli lip daemon --socket ~/.local/share/lip/lip.sock
License
Free for:
- Personal use
- Open source projects
- Startups & small businesses under $25k annual revenue
Commercial license required for organizations with $25k+ annual revenue. See pricing for Team and Enterprise plans, or LICENSE for full terms.
