Neurometric VS Code Extension

Optimize your LLM usage with cost analysis, model recommendations, and side-by-side comparisons directly in VS Code.

Features

Cost Analysis

Inline CodeLens: See cost estimates directly above LLM API calls
Analyze Current File: Get a complete cost breakdown for all detected API calls
Token Counting: Automatic token estimation for prompts

Model Recommendations

Task-Based Recommendations: Get model suggestions based on your task type
Constraint-Aware: Recommendations consider your budget and latency requirements
Alternative Options: See ranked alternatives with tradeoff explanations

Arena Mode

Side-by-Side Comparison: Compare multiple models on the same prompt
Real-Time Metrics: See cost, latency, and token usage for each response
Easy Integration: Copy responses or insert directly at cursor

MCP Server Integration

The extension includes an MCP (Model Context Protocol) server that enables AI assistants to:

Analyze costs for prompts
Get model recommendations
Compare models programmatically
Access pricing information

Installation

From VS Code Marketplace

Open VS Code
Go to Extensions (Ctrl+Shift+X / Cmd+Shift+X)
Search for "Neurometric"
Click Install

From Source

cd packages/vscode-extension
npm install
npm run compile

Then press F5 to launch the Extension Development Host.

Usage

Keyboard Shortcuts

Ctrl+Shift+N / Cmd+Shift+N: Analyze costs in current file
Ctrl+Shift+A / Cmd+Shift+A: Open Arena Mode
Ctrl+Shift+R / Cmd+Shift+R: Get model recommendation

Commands

Access via Command Palette (Ctrl+Shift+P / Cmd+Shift+P):

Neurometric: Analyze Cost
Neurometric: Open Arena Mode
Neurometric: Get Model Recommendation
Neurometric: Sign In

Right-click in the editor to:

Analyze Cost at Cursor
Compare Selection with Multiple Models

Configuration

Setting	Default	Description
`neurometric.showCostInline`	`true`	Show cost estimates via CodeLens
`neurometric.defaultModels`	`["gpt-4o", "claude-3-5-sonnet", "gemini-1.5-pro"]`	Models for Arena comparisons
`neurometric.autoAnalyze`	`false`	Auto-analyze on file save
`neurometric.costThreshold`	`0.01`	Highlight costs above this (USD)
`neurometric.apiEndpoint`	`https://api.neurometric.ai/v1`	API endpoint (enterprise/self-hosted)

Supported Languages

The extension detects LLM API calls in:

Python
TypeScript
JavaScript
TypeScript React (TSX)
JavaScript React (JSX)

Detected Providers

OpenAI (GPT-4, GPT-4o, etc.)
Anthropic (Claude 3, Claude 3.5)
Google (Gemini 1.5)
Azure OpenAI
AWS Bedrock

MCP Server

The extension registers an MCP server that AI assistants can use. Available tools:

`neurometric_analyze_cost`

Analyze cost and token usage for a prompt.

`neurometric_recommend_model`

Get model recommendations based on task type.

`neurometric_compare_models`

Compare multiple models side-by-side.

`neurometric_get_pricing`

Get current pricing information.

Development

# Install dependencies
npm install

# Compile TypeScript
npm run compile

# Watch mode
npm run watch

# Run linting
npm run lint

# Package extension
npm run package

License

MIT

Neurometric - AI Inference Optimizer

Neurometric