OpenAI .env Generator
Most AI apps use OpenAI + at least one vector DB. This generator produces a .env covering OpenAI, Anthropic, and Pinecone — the default AI stack for 2026.
OPENAI_API_KEY=sk- ANTHROPIC_API_KEY=sk-ant- OPENAI_MODEL=gpt-4o-mini PINECONE_API_KEY= PINECONE_INDEX=
Secrets regenerate on every page load. Want to combine stacks? Use the full generator.
How to use this .env in OpenAI
- 1.Generate OpenAI keys at platform.openai.com/api-keys — use a scoped key per project.
- 2.Generate Anthropic keys at console.anthropic.com → API Keys.
- 3.Paste everything into .env.local. OPENAI_MODEL is a runtime knob, not a key — you'll swap this often.
- 4.For Next.js, keep these server-only — prefixing with NEXT_PUBLIC_ would expose them in every bundle.
Gotchas
- ⚠OpenAI keys are billed per-request — set per-project spend limits in the dashboard.
- ⚠Never use OPENAI_API_KEY directly in client code. Proxy via an API route.
- ⚠Rate-limit your endpoint so a leaked key doesn't drain your budget in minutes.
Common keys explained
OPENAI_API_KEYsk-... — used by the OpenAI SDK. Server-only.
OPENAI_MODELgpt-4o-mini, gpt-4o, etc. Optional but handy for A/B tests.
ANTHROPIC_API_KEYsk-ant-... — Anthropic SDK.
PINECONE_API_KEYPinecone vector DB key.
PINECONE_INDEXPinecone index name for your embeddings.
Loading and validating the .env in OpenAI
import OpenAI from "openai";
import Anthropic from "@anthropic-ai/sdk";
// Both SDKs read their key from the environment by default:
const openai = new OpenAI(); // uses OPENAI_API_KEY
const anthropic = new Anthropic(); // uses ANTHROPIC_API_KEY
// Keep the model name in env so you can switch without a deploy
const model = process.env.OPENAI_MODEL ?? "gpt-4o-mini";
export async function summarize(text: string) {
const res = await openai.chat.completions.create({
model,
messages: [{ role: "user", content: `Summarize: ${text}` }],
});
return res.choices[0].message.content;
}OpenAI: env vars in development, staging and production
AI API keys are billed per token, so a leaked key costs money within minutes. Bots scan public GitHub for sk- and sk-ant- patterns. Use a separate key per environment and per service, set monthly spend limits in each provider's console, and revoke dev keys when a laptop is replaced.
Never call these APIs directly from the browser with your key. Route requests through your backend or an edge function that adds the key server-side and applies rate limits per user.
OpenAI: .env errors and fixes
401 Incorrect API key providedThe key was revoked, belongs to another organization or project, or has a trailing space or newline from copy-paste. Recreate it and paste it without quotes.
429 You exceeded your current quotaThis is a billing limit, not a rate limit. Add credits or raise the project's limit in the provider console. Retrying will not help.
OpenAIError: The OPENAI_API_KEY environment variable is missing or emptyThe SDK was created before dotenv loaded, or in client-side code. Load env first and keep the client on the server.
OpenAI .env FAQ
Can I use one API key for development and production?
You can, but you should not. Separate keys let you set separate spend limits and revoke a leaked dev key without breaking production.
Where do vector database keys (Pinecone, Qdrant) go?
In the same server-side .env. They give full read/write access to your index, so treat them like database passwords.
How do I check if my AI key leaked?
Run your .env through the leak checker linked below, search your git history (git log -p | grep sk-), and check the provider's usage dashboard for traffic you don't recognize.