OneAPIKey
ONEAPIKEYv1.0.0STATUSALL_SYSTEMS_NOMINALUPTIME_24H100.0%PROVIDERS12MODELS18SPREAD125xFALLBACKTIER_CASCADEBUILDBUILD 1F359TZUTCOPUS://ACTIVE
// 000ONEAPIKEY · INDEX

ONE
KEY.
EVERY
API.

[ INITIATE_SESSION ]
./browse_catalogor run npx @oneapikey/cli predict

Aggregating 18 models from 12 providers behind one key. Smart routing, fall-down tiers, real-time cost analyzer, wallet ledger, OAuth 2.0 server. Built by Opus.

In private preview: OneAPIKey Pay · Wallet · Device login and 1 more
See full status →
// 001UPSTREAM_PROVIDERS

12 PROVIDERS.
ONE CREDENTIAL.

REGISTRY · /providers12 active
[#]PROVIDERSLUGMODELSP50STATUS
01OpenAI/openai7342msOPERATIONAL
02Anthropic/anthropic5289msOPERATIONAL
03Google/google6412msOPERATIONAL
04xAI/xai2521msOPERATIONAL
05Groq/groq5 88msOPERATIONAL
06Mistral/mistral3367msOPERATIONAL
07DeepSeek/deepseek3612msOPERATIONAL
08Together AI/together4 — DEGRADED
09Perplexity/perplexity2445msOPERATIONAL
10Cohere/cohere2298msOPERATIONAL
11ElevenLabs/elevenlabs4183msOPERATIONAL
12Replicate/replicate8967msOPERATIONAL
// 002WHY_THIS_EXISTS
// 01
STOP HARDCODING

One SDK. One key. Swap models with a string. Fall-down tiers handle outages without code changes.

// 02
STOP OVERPAYING

Spread between cheapest and most expensive model is up to 125×. Optimizer surfaces switches in dollars per month.

// 03
STOP STITCHING

Wallet, OAuth, webhooks, audit log, OpenAPI spec — assembled. No 6-tool integration.

// ACTIVATE

GET A KEY. NOW.

[ ENROLL ]
// 003DROP_IN_REPLACEMENT

ZERO
REWRITE.

OpenAI SDK. Change one line — baseURL. Get every provider, every fallback, every cost optimization. Same tools, ten times the leverage.

example.tstypescript
import OpenAI from "openai";

const client = new OpenAI({
  apiKey: process.env.ONEAPIKEY_API_KEY,
  baseURL: "https://oneapikey.app/v1",
});

const res = await client.chat.completions.create({
  model: "auto/tier-cascade",     // T1 → T2 → T3
  messages: [{ role: "user", content: "ship it" }],
});

// response.headers:
//   x-oneapikey-cost-usd: 0.000341
//   x-oneapikey-cheaper-available: deepseek-v3
//   x-oneapikey-cheaper-savings-pct: 78.4
0
PROVIDERS
0
MODELS
0×
COST_SPREAD
0.0%
UPTIME_24H
OpenAI×Anthropic×Google×xAI×Groq×Mistral×DeepSeek×Together×Perplexity×Cohere×ElevenLabs×Replicate×
// 004FAQ · /var/log/objections
[01]How does your pricing compare to going direct?
→We add a small, fully transparent markup — typically 15–25% — on top of upstream pricing. In exchange you get unified billing, smart fallbacks, dashboards, and no provider registration overhead. You can see the exact markup on each API's detail page.
[02]Is it really compatible with the OpenAI SDK?
→Yes. Every language model on OneAPIKey exposes an OpenAI-compatible chat-completions endpoint. Set baseURL to ours and pass your OneAPIKey — that's it.
[03]Do you store my prompts or responses?
→We store metadata needed to bill and analyze your usage (timestamps, token counts, request IDs). Raw request and response bodies are discarded by default. You can opt into logging bodies for your own debugging.
[04]What about rate limits?
→Every key has configurable RPM/RPD limits, and we enforce concurrency per plan. If an upstream provider rate-limits us, we automatically fall back to an equivalent model from another vendor when you opt in.
[05]How are prices kept up to date?
→We rebase prices against a CPI index on a schedule you control. Upstream price changes are mirrored within 24 hours. You always see both upstream cost and the effective price you pay.
[06]Can I connect my existing app in a couple of clicks?
→Yes — our SSO flow lets external apps hand off a user to OneAPIKey, provision a key, and redirect back. Your users never paste a raw provider API key again.
2026 · BUILD 1F359 · OPUS://STUDIO

ONE KEY.
STOP STITCHING.

[ ENROLL ]