LLM Cost Tracker by PHANTSOM

Know what every LLM API call costs — before it hits your bill.

What It Does

LLM Cost Estimator scans your Python / TypeScript / JavaScript code and shows real-time inline cost estimates next to every LLM API call. No more surprise bills.

response = client.chat.completions.create(
    model="gpt-4o",          # ← 🟡 $0.003/call · ~$9.00/mo [GPT-4o]
    messages=messages,
    max_tokens=1000
)

Features

1. Inline Cost Badges

Color-coded cost decorations appear at the end of every line containing an LLM model reference:

• 🟢 Green — Very cheap (< $0.001/call)
• 🟡 Yellow — Moderate ($0.001–$0.01/call)
• 🟠 Orange — Expensive ($0.01–$0.05/call)
• 🔴 Red — Very expensive (> $0.05/call)
• 💀 Skull — Dangerously expensive (> $0.10/call)

2. Rich Hover Tooltips

Hover over any model name to see:

• Full pricing per 1M tokens (input/output/cached)
• Estimated cost for this specific call
• Daily / Weekly / Monthly / Yearly projections
• Model context window and max output info

3. CodeLens (Above Each Call)

Cost summary appears above each API call line with:

• Model name + per-call cost + monthly projection
• 💡 Cheaper alternative suggestions for expensive models

4. Status Bar

Bottom status bar shows total file cost at a glance:

• Number of LLM calls detected
• Estimated monthly cost
• ⚠️ Budget warning indicator

5. Budget Alerts

Set a monthly budget and get notified when your file's estimated costs exceed it.

6. Smart Detection

• Extracts max_tokens from nearby code for accurate output estimates
• Analyzes prompt/message strings for input token estimation
• Handles date-suffixed model IDs (e.g., gpt-4o-2024-08-06)
• Skips comments automatically

Supported Providers & Models (40+)

Supported Languages

• Python (.py)
• TypeScript (.ts)
• JavaScript (.js)
• TypeScript React (.tsx)
• JavaScript React (.jsx)

Configuration

Open VS Code Settings (Ctrl+,) and search for LLM Cost:

Commands

Installation

From VS Code Marketplace

1. Open VS Code
2. Go to Extensions (Ctrl+Shift+X)
3. Search for "LLM Cost Estimator"
4. Click Install

From VSIX (Manual)

code --install-extension llm-cost-estimator-1.0.0.vsix

Build From Source

git clone https://github.com/phantsom/llm-cost-estimator.git
cd llm-cost-estimator
npm install
npm run compile
# Press F5 in VS Code to launch Extension Development Host

How It Works

1. Detect — Regex patterns scan your code for model name strings ("gpt-4o", "claude-3-sonnet", etc.)
2. Resolve — Matches model strings against a pricing database of 40+ models
3. Estimate — Calculates cost using detected max_tokens + estimated input tokens
4. Decorate — Applies color-coded inline decorations, hover tooltips, CodeLens, and status bar
5. Alert — Warns you if estimated monthly costs exceed your budget

Why This Exists

As a Generative AI Engineer building production LLM systems, I noticed teams often have zero visibility into per-call API costs during development. A single model choice (GPT-4 vs GPT-4o-mini) can mean the difference between $50/month and $5,000/month.

This extension makes cost a first-class citizen in your development workflow.

Author

Soham Dahivalkar — Generative AI Engineer | PHANTSOM

• LinkedIn
• GitHub
• PyPI — ai-bridge-kit
• Amazon — Published Author

License

MIT — Free for personal and commercial use.

Built with 🧠 by PHANTSOM
"The Invisible Force Behind Intelligent Systems"