promptlens

Plug-in LLM cost tracking for production apps. One env var, two lines of code.

Install

npm install canopy-sdk openai

Usage

import OpenAI from "openai";
import { promptlensOpenAI } from "canopy-sdk";

const openai = new OpenAI(
  promptlensOpenAI({
    apiKey: process.env.OPENAI_API_KEY!,
    feature: "my-feature", // optional — appears in dashboard
  })
);

// All completions are logged automatically
await openai.chat.completions.create({
  model: "gpt-4o-mini",
  messages: [{ role: "user", content: "Hello" }],
});

# Only new secret — from Canopy dashboard → API Keys
export PROMPTLENS_KEY=pl_live_...

Anthropic

import Anthropic from "@anthropic-ai/sdk";
import { promptlensAnthropic } from "canopy-sdk";

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

Self-hosted proxy

promptlensOpenAI({
  apiKey: process.env.OPENAI_API_KEY!,
  baseUrl: "https://promptlens.mycompany.com",
});

Or set PROMPTLENS_BASE_URL in the environment.

Lower-level API

If you prefer spreading config yourself:

import { createOpenAIConfig, resolvePromptLensConfig } from "canopy-sdk";

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

vs TokenGuard

TokenGuard is a browser extension that optimizes chat usage. Canopy is for teams that call OpenAI/Anthropic from code and need spend broken down by feature and prompt template.