DaoXECalling several model vendors directly means separate accounts, separate cards, and separate bills — and in many regions the official checkout simply fails. A gateway collapses that into one base URL (https://daoxe.com/v1), one key, and one bill across GPT · Claude · Gemini · DeepSeek · Kimi · Qwen · Doubao. This guide shows how to use DaoXE well, honestly, and how to verify it.
Why a gateway — and why this one#
One key, many models
GPT, Claude, Gemini, DeepSeek, Kimi and more behind a single endpoint and one bill. No per-vendor signup, cards, or reconciliation.
Three protocols, natively
OpenAI Chat Completions, OpenAI Responses, and Anthropic Messages — so Claude Code works properly, not just the OpenAI shape.
Honest by design
No silent bait-and-switch. Account-scoped model list, transparent usage, and a method to catch downgrading.
Verifiable, not trusted
Point the open benchmark at DaoXE and at anyone else. Re-run it yourself; the numbers are reproducible.
Drop-in for your tools
Change one base URL in Cursor, Cline, Aider, LobeChat and OpenAI/Anthropic SDKs. Existing code keeps working.
Built for overseas devs
For developers abroad, overseas Chinese communities, and Russian-speaking developers. Not available in mainland China.
Start in three steps#
- Get a key. Create an account and an API key in your DaoXE dashboard.
- Discover your model IDs. Model IDs are account-scoped — call the models endpoint and copy an exact ID. Don't hardcode a name from a blog post.
- Point your client at the base URL. Set the base URL to
https://daoxe.com/v1, paste the key, and use an exact model ID. For Claude Code, use the native Messages path.
export DAOXE_API_KEY="your_api_key"
curl --fail-with-body --show-error --silent \
https://daoxe.com/v1/models \
-H "Authorization: Bearer ${DAOXE_API_KEY}"Three protocols, one endpoint#
| Protocol | Endpoint | Use it for |
|---|---|---|
| OpenAI Chat Completions | POST /v1/chat/completions | Most OpenAI-compatible clients & SDKs |
| OpenAI Responses | POST /v1/responses | Projects on the Responses API |
| Anthropic Messages (Claude) | POST /v1/messages | Claude Code, Anthropic SDKs, Messages-native tools |
| Image generation | OpenAI-compatible images | OpenAI-style image endpoints where available |
Don't trust us — benchmark us#
“It's fast and reliable” is a claim; a number you can re-run is proof. The companion project llm-gateway-benchmark probes a set of models and reports success rate, p50/p95 latency, and price per 1M tokens. It never stores your key, prompt, or responses. Read the gateway benchmark guide to run it.
Verification is the whole point
Point the CLI at DaoXE, at the official API, and at your current provider, then compare. If a provider gets cagey when you ask to benchmark it, that's your answer. Learn how to detect model swapping.
Payment & real cost#
Alipay / WeChat Pay / USDT — a flat 1 RMB = $1
Every method tops up at a flat 1 RMB = $1 of credit, while models bill at their official USD price. At a real ~7 RMB/USD rate, Western flagships (GPT / Claude / Gemini / Grok) come out around 1/7 of official cost (≈85% off) — and because USDT tops up the same way, USD holders abroad get the same deal. Honest caveat: cheaper domestic models (DeepSeek / Qwen / GLM) still beat going direct, just by less — it is not “everything is 1/7”. Rates move; daoxe.com/pricing is authoritative, and a small top-up verifies it.
Pick a guide#
Cheapest Claude / GPT / Gemini API
Where real savings come from — and the one trap to avoid.
DaoXE vs official API cost
How gateway pricing compares to official APIs, honestly.
Cheaper endpoint in Cursor & Claude Code
Exact setup, the honest caveats, and how to verify.
OpenRouter alternative
A multi-model gateway you can benchmark — with native Claude protocol.
LLM gateway price & latency benchmark
Measure price, latency and success rate yourself.
Detect relay model swapping
Catch downgrading, quantization and context truncation in any proxy.
Availability & honesty
DaoXE is not available in mainland China; it targets developers overseas, overseas Chinese communities, and Russian-speaking developers. Pricing, free credit and uptime are set by the service — this guide never invents numbers. Current price advantage: typically ~30–80% below list via low-price groups (see the pricing page).
Frequently asked questions#
Is DaoXE OpenAI-compatible?
Yes. Set the base URL to https://daoxe.com/v1 and use your key; existing OpenAI SDK code runs unchanged. It also exposes OpenAI Responses and native Anthropic Messages.
Does Claude Code work through DaoXE?
Yes — it supports the Anthropic Messages protocol natively at /v1/messages, so tool use and streaming behave correctly. See the Cursor & Claude Code guide.
Which model ID should I use?
Always an exact ID from GET /v1/models for your account — IDs are account-scoped, so don't copy strings from other people's examples.
How do I know I'm getting the real model?
Run the open, reproducible benchmark against DaoXE and the official API and compare. See how to detect model swapping.
Is there a free trial or credit?
Free credit on signup is to be confirmed. Check the current offer on the official site.
Try DaoXE — and benchmark it yourself
One key for GPT, Claude, Gemini, DeepSeek and more. Point the open benchmark at us and compare — don't take our word for it.