Cut AI coding costs 58% with pre-indexed codebase context
this week
discussions
this week
vexp is a local-first context engine designed to improve how AI coding agents gather code context. Instead of letting an agent scan large parts of a repository and consume a big context window, vexp builds a dependency graph of your code (functions, classes, types and their relationships) and uses it to assemble a smaller, task-relevant context bundle. It returns key “pivot” nodes with full source and surrounding code as skeletons (signatures, docstrings, and types) to cut token usage. vexp runs entirely on your machine with a local SQLite index and file watching, and it’s delivered as a VS Code extension and a CLI that integrates with multiple MCP-compatible agents. It also captures and recalls session memory from tool calls, and can show callers/dependents to understand change impact before refactoring. The product emphasizes zero network calls and no account requirement.
AI coding agents waste tokens by loading irrelevant code context.
Local dependency-graph context + memory returns only needed code.
Developers using AI coding agents in editors.
Promo codes available for this launch.
PRODUCTHUNT
1 month free of PRO
Expires Mar 16, 2026
Share your thoughts about this tool.
Sign in to leave a comment.
No comments yet. Be the first to leave one.