DEV Community: ZeroShell

One API Key for Claude, GPT-5, and Gemini — A Practical Relay Setup for Cursor, ChatBox, and Claude Code

ZeroShell — Thu, 05 Nov 2026 03:18:14 +0000

Disclosure: I run ZeroShell, the relay used in this article. Everything below — prices, model lists, the group pitfall — is verifiable on the public pages linked at the end.

Most developers I know keep three or four separate API accounts: one for Claude, one for GPT, one for Gemini, plus a DeepSeek key for cheap experiments. Four balances, four dashboards, four chances to run dry mid-task.

A relay endpoint collapses that into one account. This is the setup I give people who ask, including the one mistake that wastes everyone's first ten minutes.

The endpoint

ZeroShell is an OpenAI-compatible relay:

Base URL: https://zeroshell.top/v1
Anthropic-compatible base: https://zeroshell.top (native /v1/messages)
46 models across Claude, GPT-5, Gemini, DeepSeek, Grok, Kimi, Qwen, and GLM families, all on one balance.

The pitfall: model catalogs are per-group

This is the part that bites. On ZeroShell, API keys belong to a group, and each group sees a different model catalog:

The budget group sees 27 models: DeepSeek, GLM, GPT, Grok, Kimi, Qwen — no Claude.
The official and stable groups carry the Claude lineup (22 models in the official group).

If you create a budget-group key and ask for claude-sonnet-4-6, you'll get a "model not found" error that looks exactly like a typo. It isn't. Check which group your key is in before you debug anything else.

What it costs

All prices are public (no login needed) and listed per million tokens. A few current examples:

Model	Input	Output
gpt-5.6-luna	¥0.075	—
gpt-5.6-sol	¥0.375	¥2.25
claude-sonnet-4-6	¥2.025	¥10.125
gemini-3.1-pro-preview	¥0.75	¥4.5

Top-ups are 1:1 — ¥1 buys $1 of quota — with a ¥1 minimum. New accounts get ¥1 of credit after email verification, and an invite link gives ¥1 to both sides, so you can verify your whole toolchain before spending anything.

Two honest caveats:

Top-ups are Alipay-only right now. If you can't use Alipay, this station isn't practical for you yet.
Relay ≠ official. You're trusting an operator with routing and billing. Judge any relay — this one included — by whether prices are public, whether you can start with a tiny top-up, and whether the models behave like the real thing on your own test prompts.

Point your tools at it

Cursor — Settings → Models: add your ZeroShell key as the OpenAI API key and override the OpenAI base URL with https://zeroshell.top/v1. Pick a model your key's group actually contains (see the pitfall above).

ChatBox / any OpenAI-compatible client — choose the OpenAI API provider, set the API host to https://zeroshell.top, path /v1, and paste your key.

Claude Code — create an official-group (or stable-group) key, then set ANTHROPIC_BASE_URL=https://zeroshell.top and ANTHROPIC_API_KEY to your key. Claude models answer over the native Messages API, so Claude Code works unmodified. I run my own local Claude Code through this exact setup.

Verify before you commit

The workflow I'd recommend with any relay:

Sign up, take the free ¥1 credit.
Run three prompts you know well — one reasoning, one coding, one long-context — and compare against what you expect from the genuine model.
Top up the minimum (¥1) only after the outputs check out.

If a relay is going to water models down or disappear, small tests and small top-ups are how you find out cheaply.

Links: zeroshell.top · public pricing at zeroshell.top/pricing · questions: support@zeroshell.top

Run Claude Code on a Relay Endpoint: Setup, the Model-Catalog Pitfall, and What It Actually Costs

ZeroShell — Wed, 04 Nov 2026 15:52:16 +0000

Most AI coding tools speak one of two dialects: the OpenAI chat-completions format, or Anthropic's Messages format. If you want to run those tools against a third-party relay endpoint instead of a first-party account — for cost reasons, payment reasons, or because one key should cover several model families — the setup is short. The part that wastes people's time is a catalog mismatch that produces a confusing model_not_found error. This guide covers the configuration, that pitfall, and a cost breakdown with real listed prices, using one relay (ZeroShell, which I operate — disclosure at the end) as the worked example. The steps generalize to any OpenAI-compatible relay.

The two endpoints you need to know

A relay typically exposes:

POST /v1/chat/completions — the OpenAI-compatible endpoint. Base URL form: https://zeroshell.top/v1.
POST /v1/messages — the Anthropic-native endpoint. This is the one Claude Code actually requires.

That second endpoint is the real compatibility test. Plenty of relays advertise "Claude support" but only translate OpenAI-format requests; Claude Code sends Anthropic-format requests and fails against those. Before paying anything, verify the relay documents and serves /v1/messages. On ZeroShell both formats are served, and both have been exercised with live requests — Claude models respond over the OpenAI format and over /v1/messages.

Claude Code configuration

Claude Code reads its endpoint from environment variables:

export ANTHROPIC_BASE_URL=https://zeroshell.top
export ANTHROPIC_API_KEY=<your key>
claude

Note the base URL has no /v1 suffix here — Claude Code appends /v1/messages itself. Model names come from the relay's own catalog, for example claude-sonnet-4-5, so copy them from the relay's model list rather than typing from memory; a wrong suffix is the most common self-inflicted error.

The pitfall: one account, several catalogs

Here is the part most relay documentation glosses over. On ZeroShell, an API key belongs to a group, and each group has its own model catalog:

A low-cost group key sees 27 models: DeepSeek, GLM, GPT, Grok, Kimi, and Qwen families. No Claude models at all.
An official group key sees 22 models, including the full Claude lineup.

The site advertises 46 models — that figure is the aggregate across all groups, not what any single key can call. If you create a low-cost key and point Claude Code at it, every request returns model_not_found, and the error message blames the model rather than your key's group. The fix takes thirty seconds (create a key in the correct group), but only once you know groups exist. One account can hold keys in several groups, and the balance is shared, so the practical setup is: an official-group key for Claude Code, a low-cost key for bulk OpenAI-format work.

When evaluating any relay, ask this question first: is the advertised model count per-key or per-account? The answer determines whether your first integration attempt works.

Other OpenAI-compatible tools

For tools like Continue, Aider (in OpenAI mode), or chat clients such as Cherry Studio and NextChat, configuration is the standard trio:

Base URL: https://zeroshell.top/v1
API key: a key from the group that contains your model
Model: the exact catalog name, copied, not recalled

A 30-second smoke test before wiring up any tool:

curl https://zeroshell.top/v1/chat/completions \
  -H "Authorization: Bearer <your key>" \
  -H "Content-Type: application/json" \
  -d '{"model":"gpt-5-mini","messages":[{"role":"user","content":"ping"}]}'

If this returns a completion, every OpenAI-compatible client will work. If it returns a model error, suspect the group/catalog mismatch above before suspecting the endpoint.

What it costs

Relay pricing is per million tokens, published on the pricing page. Current listed prices:

Model	Input / 1M	Output / 1M
gpt-5-nano	$0.03	$0.24
gpt-5-mini	$0.25	$2.00
gemini-2.5-pro	$1.25	$10.00
claude-sonnet-4-5	$3.00	$15.00

Worked example: a heavy personal month of 20M input and 5M output tokens costs $19.25 on gpt-5-mini, $85 on gemini-2.5-pro, and $135 on claude-sonnet-4-5. That spread is exactly why the group system exists — route summarization and boilerplate to the cheap catalog, and reserve the Claude catalog for work where its output quality earns the difference.

Getting started and honest limits

Signup is a username, a password, and an email verification code — about a minute — and new accounts receive ¥1 of credit, which is enough to run every step in this guide. Top-ups are Alipay only, starting at $1, with ¥1 = $1 of quota; invites give ¥1 to both sides. Two limits worth stating plainly: there is no card or PayPal payment yet, so this suits users who can pay with Alipay; and it is a small, new service, so treat the free credit as a way to verify your own workflow before committing real volume. Questions and failure reports go to support@zeroshell.top

Disclosure: I operate ZeroShell. Every price and behavior above is from the public pricing page and documented, tested endpoint behavior as of November 2026; check the pricing page for current figures.