aitoken-cli

Track every dollar you spend on AI APIs

Local-first • Privacy-focused • Multi-provider • Fast

---

The Problem

You're spending $200+/month on GPT-4, Claude, and other AI APIs.

But you have no idea where it's going.

The Solution

Track every API call locally. See exactly what you're spending.

``

bash

Track usage

at add -p openai -m gpt-4o -i 1500 -o 800



View stats

at stats

Total: $34.56 (245,890 tokens)





Features



- Automatic Tracking - Built-in wrappers, middleware, and SDK extensions (NEW in v1.1.0)

- Multi-Provider Support - OpenAI, Anthropic, Google, Azure, Cohere

- Automatic Cost Calculation - Up-to-date pricing for all major models

- Beautiful Stats - See usage by provider, model, and time period

- Local Storage - All data stored locally in SQLite (privacy-first)

- Fast & Lightweight - CLI tool, no GUI overhead

- JSON Export - Pipe data to other tools

- Programmatic API - Use in your code without CLI calls



Installation

bash

npm install -g aitoken-cli





CLI Usage



$3

bash

Basic usage

at add -p openai -m gpt-4o -i 1500 -o 800



With notes

at add -p anthropic -m claude-3.5-sonnet -i 2000 -o 1200 -n "Code review session"



Options:

-p, --provider   Provider (openai, anthropic, google, azure, cohere)

-m, --model      Model name

-i, --input      Input/prompt tokens

-o, --output     Output/completion tokens

-n, --notes      Optional notes

$3

bash

List recent usage (default: last 20)

at list



Filter by provider

at list -p openai



Limit results

at list -l 50



Export as JSON

at list --json > usage.json

$3

bash

Overall stats

at stats



Last 7 days

at stats -d 7



Specific provider

at stats -p anthropic



Export as JSON

at stats --json

$3

bash

Quick view of today

at today



As JSON

at today --json

$3

bash

List all models and pricing

at models



Filter by provider

at models -p openai

$3

bash

Clear all data (requires confirmation)

at clear --yes



Clear specific provider

at clear -p openai --yes



Clear before date

at clear --before 2026-01-01 --yes





Example Output

bash

$ at stats



Overall Stats



┌─────────────────┬──────────┐

│ Metric          │ Value    │

├─────────────────┼──────────┤

│ Total Requests  │ 127      │

│ Total Tokens    │ 3.2M     │

│ Total Cost      │ $24.5670 │

└─────────────────┴──────────┘



By Provider



┌──────────┬──────────┬─────────┬──────────┬────────────┐

│ Provider │ Requests │ Tokens  │ Cost     │ % of Total │

├──────────┼──────────┼─────────┼──────────┼────────────┤

│ openai   │ 85       │ 2.1M    │ $18.2340 │ 74.2%      │

│ anthropic│ 32       │ 890K    │ $5.1230  │ 20.9%      │

│ google   │ 10       │ 210K    │ $1.2100  │ 4.9%       │

└──────────┴──────────┴─────────┴──────────┴────────────┘





Automatic Tracking (v1.1.0+)



The package includes built-in automatic tracking functions - no need for exec() or CLI calls in your code!



$3

typescript

//  Method 1: Wrapper Functions (Easiest)

import { trackedGPT, trackedClaude, trackedGemini } from 'aitoken-cli/wrappers';

import OpenAI from 'openai';

import Anthropic from '@anthropic-ai/sdk';



const openai = new OpenAI({ apiKey: process.env.OPENAI_API_KEY });

const anthropic = new Anthropic({ apiKey: process.env.ANTHROPIC_API_KEY });



// Just use trackedGPT() instead of openai.chat.completions.create()

const response = await trackedGPT(openai, {

  model: 'gpt-4o',

  messages: [{ role: 'user', content: 'Hello!' }]

  // Automatically tracked! 

});



const claudeResponse = await trackedClaude(anthropic, {

  model: 'claude-sonnet-4.5',

  max_tokens: 1024,

  messages: [{ role: 'user', content: 'Explain TypeScript' }]

  // Automatically tracked! 

});

typescript

//  Method 2: Middleware Pattern (Zero Code Changes)

import { createTrackedClient } from 'aitoken-cli/middleware';

import OpenAI from 'openai';



const openai = new OpenAI({ apiKey: process.env.OPENAI_API_KEY });



// Wrap your client once

const trackedOpenAI = createTrackedClient(openai, {

  provider: 'openai',

  model: 'gpt-4o'

});



// Use it exactly like normal - tracking happens automatically

const response = await trackedOpenAI.chat.completions.create({

  model: 'gpt-4o',

  messages: [{ role: 'user', content: 'Hello!' }]

});

// Automatically tracked with zero code changes!

typescript

//  Method 3: SDK Extensions (Drop-in Replacement)

import { TrackedOpenAI, TrackedAnthropic } from 'aitoken-cli/extensions';



// Just change the import - everything else stays the same!

const openai = new TrackedOpenAI({

  apiKey: process.env.OPENAI_API_KEY

});



const response = await openai.chat.completions.create({

  model: 'gpt-4o',

  messages: [{ role: 'user', content: 'Hello!' }]

});

// Automatically tracked!





See EXAMPLES.md for 12 complete examples including Express.js, Next.js, and chatbot integrations.



Programmatic API

typescript

import { addUsage, getUsage, getStats, calculateCost } from 'aitoken-cli';



// Calculate cost for a request

const cost = calculateCost('openai', 'gpt-4o', 1500, 800);

console.log(

Cost: $${cost.toFixed(4)}

); // $0.0195



// Add a usage entry

addUsage({

  provider: 'openai',

  model: 'gpt-4o',

  promptTokens: 1500,

  completionTokens: 800,

  totalTokens: 2300,

  cost,

  timestamp: new Date().toISOString(),

  notes: 'My API call',

});



// Get usage entries

const usage = getUsage({ provider: 'openai', limit: 10 });



// Get statistics

const stats = getStats({ provider: 'openai' });

console.log(

Total: $${stats.totalCost.toFixed(2)} (${stats.totalRequests} requests)

);





Data Storage



All data is stored locally in

~/.token-tracker/usage.db

 (SQLite).



No data is sent anywhere. 100% local.



Supported Providers & Models (42 Models)



$3

- GPT-5.2, GPT-5.2 Pro, GPT-5 Mini

- GPT-4.1, GPT-4.1 Mini, GPT-4.1 Nano

- o4-mini

- GPT-4, GPT-4-32K, GPT-4 Turbo, GPT-4o, GPT-4o Mini

- GPT-3.5 Turbo, GPT-3.5 Turbo 16K

- o1-preview, o1-mini



$3

- Claude Sonnet 4.5, Claude Haiku 4.5, Claude Opus 4.5

- Claude 3.5 Sonnet, Claude 3.5 Haiku

- Claude 3 Opus, Claude 3 Sonnet, Claude 3 Haiku

- Claude 2.1, Claude 2.0, Claude Instant



$3

- Gemini 1.5 Pro, Gemini 1.5 Flash

- Gemini 1.0 Pro, Gemini Pro, Gemini Pro Vision



$3

- GPT-4, GPT-4-32K, GPT-3.5 Turbo



$3

- Command, Command Light, Command R, Command R+



Development

bash

Clone repo

git clone https://github.com/brian-mwirigi/aitoken-cli.git

cd aitoken-cli



Install dependencies

npm install



Run in dev mode

npm run dev -- add -p openai -m gpt-4o -i 1000 -o 500



Build

npm run build



Test locally

npm link

at stats

Contributing

Contributions welcome! Please open an issue or PR.

License

MIT

Author

Built by Brian Mwirigi