dev-agent
Local semantic code search for Cursor and Claude Code via MCP.
Get Started · View on GitHubÂ
v0.5.0 coming soon — Arrow functions, React hooks, and exported constants now indexed. See what’s new →
Built by engineers, for engineers. An MCP server that gives your AI tools semantic code search and context bundling. Savings scale with task complexity — up to 42% on debugging tasks.
Why it saves money
dev-agent doesn’t just search — it bundles context so Claude reads less:
| What dev-agent does | Manual equivalent | Savings |
|---|---|---|
| Returns code snippets in search | Read entire files | 99% fewer input tokens |
dev_plan bundles issue + code + commits | 5-10 separate tool calls | 29% cost reduction |
| Semantic search finds relevant code | grep chains + manual filtering | 42% cost reduction |
The harder the task, the bigger the savings.
Measured results by task type
| Task Type | Cost Savings | Time Savings | Why |
|---|---|---|---|
| Debugging | 42% | 37% | Semantic search beats grep for “where is the bug?” |
| Exploration | 44% | 19% | Find code by meaning, not keywords |
| Implementation | 29% | 22% | dev_plan bundles context in one call |
| Simple lookup | ~0% | ~0% | Both approaches are fast |
Trade-off: Faster but sometimes less thorough. Baseline Claude provides more diagnostic shell commands. dev-agent excels when you need to explore or understand code.
Same question, different approach
We asked Claude Code: “Debug why search returns duplicates”
Without dev-agent
Claude’s approach:
- Read indexer/index.ts (441 lines)
- Read vector/store.ts (258 lines)
- Read scanner/typescript.ts (full file)
- Read scanner/markdown.ts (full file)
Result: 18+ tool calls, 10 files read → $1.37, 12 minutes
18,800 input tokens consumed
Same root cause identified. 42% cheaper. 37% faster.
Context bundling: dev_plan
For implementation tasks, dev_plan bundles everything in one call:
Without dev-agent
Claude’s approach for “Implement issue #61”:
gh issue view 61 --json title,body # Fetch issue
grep "--json" -r packages/cli # Find existing flags
Read search.ts # Check implementation
Read mcp.ts # Check implementation
Read config.ts # Check file writes
# ... 5+ more tool callsResult: $0.55, 5.7 minutes
How it works
Key insight: dev-agent returns code snippets with context — Claude doesn’t read entire files. This is why input tokens drop by 99%.
Quick Start
Install
npm install -g dev-agentIndex your repository
cd your-project
dev index .Connect to your AI tool
dev mcp install --cursor # For Cursor
dev mcp install # For Claude Code9 MCP Tools
| Tool | What it does |
|---|---|
dev_search | Semantic code search — returns snippets, not just paths |
dev_plan | Context bundling — issue + code + commits in one call |
dev_refs | Find callers/callees of any function |
dev_map | Codebase structure with change frequency |
dev_history | Semantic search over git commits |
dev_explore | Find similar code, trace relationships |
dev_gh | Search GitHub issues/PRs semantically |
dev_status | Repository indexing status |
dev_health | Server health checks |
When to use it
| Scenario | dev-agent? | Expected Savings |
|---|---|---|
| Debugging unfamiliar code | âś… Yes | 42% cost |
| Exploring large codebase | âś… Yes | 44% cost |
| Implementing GitHub issues | âś… Yes | 29% cost |
| Small codebase you know | ❌ Skip | ~0% |
| Need exhaustive file reads | ⚠️ Maybe | Trade speed for thoroughness |
Features
- Context Bundling —
dev_planreplaces 5-10 tool calls with one - Code Snippets — Search returns code, not just file paths
- 100% Local — Your code never leaves your machine
- Sub-second Search — Fast even on large repos with LanceDB
- 1379+ Tests — Production-grade reliability
Read the story behind dev-agent →Â
MIT License • Built by prosdevÂ