Skip to content

Claws-ZH/awesome-ai-api

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

40 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

awesome-ai-api

The world's largest open-source hub for AI API gateways & reseller reviews.
Curated · community-driven · fully transparent · every gateway probed daily.

Awesome License: MIT PRs Welcome Total API verified Updated

🇨🇳 中文 · English · 🌐 Live site


What makes us different

We don't just list gateways — we probe every one of them daily for a real OpenAI-compatible /v1/models endpoint. No 200 OK on the homepage inflation.

  • 🔌 Ground-truth API detection — every 🔌 badge = endpoint confirmed live
  • 🔍 Transparency — no paid placement; raw data is Git-history-tracked
  • 🌏 Global — bilingual (EN/中文), covers Western + Chinese markets
  • 🛠 Developer-first — structured JSON, daily history snapshots, PR-editable
  • 🤝 Community-driven — pricing & reviews by people who actually pay for these services

🎯 Quick Picks

Confused by 200+ gateways? Start here.

Use case Recommendation Why
🌐 Global, one API for all models OpenRouter Pioneer of the pattern; most stable, highest fees
👨‍💻 Best Claude Code in China PackyAPI Claude-Code native, active upstream, ¥1/sample
🏢 Enterprise / invoices 柏拉图AI · 云雾 Long-running, multi-region, invoices, DeepSeek/MidJourney too
💰 Cheap Claude (accept risk) Terminal.Pub · XcodeBest ~0.15% of official; new = verify before topping up
🆓 Free trial credit AnyRouter · 发现AI Credits on signup, no card
🏠 Self-hosted One API · New API OSS; the engine 60% of the CN list runs on
Inference-only (Llama/Qwen/DeepSeek) Groq · Together · Fireworks Not resellers; own GPUs; US cards only
🌯 Multi-modal (video/music/image) 神马AI · 302.AI MidJourney, Kling, Suno, PPT generators

⚠️ Never pre-pay large amounts. This industry has weekly rug-pulls. Top up ¥10 first, validate quality, then consider going up.


💚 Support this project

This repo is free, ad-free, and has no paid placement. If the leaderboard saved you from topping up a rug-pull, consider chipping in:

  • USDT (any EVM chain — ERC20 / BSC / Polygon / Arbitrum / Base / Optimism):
    0xa5c74e7D3f0c8f1c0d7A395A6B7861Ab0A64cA7F
  • Star the repo — also helps a lot; boosts us in GitHub trending.
  • 📝 Open a PR with a gateway report — the most valuable contribution.

💡 Double-check the chain before sending. Sending from an exchange? Withdraw on ERC20 / BSC / Polygon / Arbitrum / Base / Optimism. Do NOT use TRC20 (TRON) — that address doesn't exist on Tron.


Table of Contents


🏆 Full Leaderboard

Auto-generated daily at 10:00 SGT (UTC+8) from live probes. See methodology · history · raw data.

🔌 = confirmed /v1/models endpoint (real API, not just a marketing page).

📊 Click to expand the full leaderboard

Last updated: 2026-05-08 12:53 (SGT)

Total: 206 gateways · 🔌 101 with confirmed /v1/models endpoint · 🟢 127 Verified · 🟡 8 Probable · 🧰 8 OSS · 🔍 63 Needs review

Top engines detected: new-api × 38 · one-api × 11 · dify × 8 · litellm × 4 · openrouter × 2

# Gateway Region API Models Engine Payment Score Latency Tier
🥇 api.vveai.com cn 🔌 claude, gpt, gemini 9.9 193 ms 🟢 Verified
🥈 api.v36.cm cn 🔌 claude, gpt, gemini 9.9 223 ms 🟢 Verified
🥉 api.v3.cm cn 🔌 claude, gpt, gemini 9.9 243 ms 🟢 Verified
4 api.oneabc.org cn 🔌 claude, gpt, gemini 9.9 263 ms 🟢 Verified
5 Flow2API cn 🔌 claude, gpt, gemini new-api 9.9 273 ms 🟢 Verified
6 ePhone AI cn 🔌 claude, gpt, gemini 9.9 316 ms 🟢 Verified
7 apipro.maynor1024.live cn 🔌 claude, gpt, gemini 9.9 322 ms 🟢 Verified
8 ClaudeCN cn 🔌 claude, gpt, gemini card 9.9 332 ms 🟢 Verified
9 Avian global 🔌 8 models 9.9 346 ms 🟢 Verified
10 api.gpt.ge cn 🔌 claude, gpt, gemini 9.9 366 ms 🟢 Verified
11 AIHubMix global 🔌 234 models openrouter 9.9 391 ms 🟢 Verified
12 PackyAPI (PackyCode) cn 🔌 claude, gpt, gemini wechat 9.9 412 ms 🟢 Verified
13 TiMi CC global 🔌 gpt, qwen wechat 9.9 526 ms 🟢 Verified
14 api-gptgod-work global 🔌 345 models 9.9 541 ms 🟢 Verified
15 AnyRouter global 🔌 22 models one-api 9.9 544 ms 🟢 Verified
16 毫秒API MJ gpt claude AI luma deepseek 中转 AI中转 chatgpt中转 claud cn 🔌 claude, gpt, chatgpt 9.9 690 ms 🟢 Verified
17 jeniya.cn cn 🔌 claude, gpt, chatgpt 9.9 802 ms 🟢 Verified
18 Doro AI cn 🔌 claude, gpt, gemini new-api 9.9 1473 ms 🟢 Verified
19 Claude Code/GPT/Gemini API 中转代理 cn 🔌 claude, gpt, gemini wechat 9.9 1719 ms 🟢 Verified
20 便携AI聚合API cn 🔌 claude, gemini, openai new-api 9.9 2088 ms 🟢 Verified
21 API Market global 🔌 gpt, gemini, anthropic 9.9 2403 ms 🟢 Verified
22 onetoken.one cn 🔌 claude, gemini, openai new-api 9.8 176 ms 🟢 Verified
23 api.onechats.top cn 🔌 claude, gemini, openai new-api 9.8 220 ms 🟢 Verified
24 chatapi.onechats.top cn 🔌 claude, gemini, openai new-api 9.8 222 ms 🟢 Verified
25 api-deerapi-com cn 🔌 claude, gemini, openai 9.8 224 ms 🟢 Verified
26 KKSJ AI模型中转API cn 🔌 claude, gemini, openai new-api 9.8 226 ms 🟢 Verified
27 api.dzzi.ai cn 🔌 claude, gemini, openai new-api 9.8 229 ms 🟢 Verified
28 api.ekan8.com cn 🔌 claude, gemini, openai new-api 9.8 256 ms 🟢 Verified
29 35.aigcbest.top cn 🔌 claude, gemini, openai new-api 9.8 260 ms 🟢 Verified
30 api-gueai-com cn 🔌 claude, gemini, openai 9.8 278 ms 🟢 Verified
31 api.gemai.cc cn 🔌 claude, gemini, openai new-api 9.8 281 ms 🟢 Verified
32 Aiberm global 🔌 claude, gemini, openai 9.8 284 ms 🟢 Verified
33 openclaudecode.cn cn 🔌 claude, gemini, openai new-api 9.8 285 ms 🟢 Verified
34 aigcbest.top cn 🔌 claude, gemini, openai new-api 9.8 339 ms 🟢 Verified
35 apirouter.ai cn 🔌 claude, gemini, openai new-api 9.8 355 ms 🟢 Verified
36 api.openai-ch.top cn 🔌 claude, gemini, openai new-api 9.8 364 ms 🟢 Verified
37 api.kissapi.ai cn 🔌 claude, gemini, openai new-api 9.8 389 ms 🟢 Verified
38 api.cursorai.art global 🔌 claude, gpt, chatgpt 9.8 390 ms 🟢 Verified
39 api.soruxgpt.com cn 🔌 claude, gemini, openai new-api 9.8 425 ms 🟢 Verified
40 api-chatfire-cn cn 🔌 claude, gemini, openai 9.8 430 ms 🟢 Verified
41 xcode.best cn 🔌 claude, gemini, openai new-api 9.8 445 ms 🟢 Verified
42 api.ifopen.ai cn 🔌 claude, gemini, openai new-api 9.8 448 ms 🟢 Verified
43 api.ikuncode.cc cn 🔌 claude, gemini, openai new-api 9.8 521 ms 🟢 Verified
44 azapi-com-cn cn 🔌 claude, gemini, openai 9.8 656 ms 🟢 Verified
45 api-mnapi-com cn 🔌 claude, gemini, openai 9.8 671 ms 🟢 Verified
46 terminal.pub cn 🔌 claude, gemini, openai new-api 9.8 680 ms 🟢 Verified
47 api-996444-cn cn 🔌 claude, gemini, openai 9.8 684 ms 🟢 Verified
48 柏拉图AI_API中转站 (api.bltcy.top) cn 🔌 claude, gpt, openai 9.8 692 ms 🟢 Verified
49 chintao.cn cn 🔌 claude, gemini, openai new-api 9.8 799 ms 🟢 Verified
50 duckcoding.ai cn 🔌 claude, gemini, openai new-api 9.8 844 ms 🟢 Verified

Top 50 shown. See data/_leaderboard.md for the full list of 206 gateways.

📌 Want your gateway listed? Open a PR with a filled gateway template. We accept any provider that meets our listing criteria.


❓ FAQ

What exactly is an "AI API gateway" or "中转站"?

A third-party service that exposes Claude, GPT, Gemini, DeepSeek, Qwen, and other frontier LLMs through a single OpenAI-compatible HTTP API. Most of them sit in front of the official providers and resell capacity, often at discounted prices, with friendlier payment options (WeChat/Alipay), or with extra features (rate-limit pooling, routing, logging, invoices).

How is this different from OpenRouter's /models page, helpaio.com, or apicompare.best?
  • OpenRouter only lists gateways running inside OpenRouter's own routing graph. We list every relay we can find, including the dozens of small CN players.
  • helpaio.com is a blog with human-written reviews. We publish machine-verified data (live /v1/models probes, engine fingerprints, uptime) alongside reviews.
  • apicompare.best compares prices. We compare reachability, authenticity, and stability — cheaper doesn't help if the gateway disappears next week.
How do you decide the leaderboard score?

We combine: reachability (HTTP 2xx), confirmed /v1/models endpoint, real-model count, model-keyword density on the landing page, latency, and 30-day uptime from our daily snapshots. The formula lives in scripts/generate_leaderboard.py — no black box.

Is this a paid directory? Who pays you to be listed?

No one. There is zero paid placement. Listings are either auto-discovered or PR'd by the community. The repo is MIT-licensed and ad-free. If the leaderboard helped you, consider supporting the project.

Can I trust the top-ranked gateways with large amounts of money?

No. This industry sees weekly rug-pulls. A high score means a gateway is currently reachable, has a real API, and has been stable for 30 days — not that it is solvent or will still be online next month. Always top up small amounts first (≤ US$2), validate model quality, and scale slowly.

I got scammed by a gateway. How do I report it?

Open a GitHub Discussion under the Reports category with public evidence (screenshots, transaction IDs, Telegram group links). Two independent reports + a 14-day operator response window = the gateway lands on data/blacklist.json.

How do I add my gateway?

Add the URL to data/candidates.txt, then (optionally) add a human-curated name, region, and verdict to data/sites.yaml. Open a PR. The daily cron will pick it up.

Do you support non-OpenAI-compatible APIs (raw Anthropic, raw Gemini)?

We primarily probe /v1/models (OpenAI-compatible), because it's the lingua franca. Gateways that only expose raw Anthropic or Gemini endpoints can still be listed via data/sites.yaml with a verdict: likely_relay override and a note explaining the protocol.


📊 Categories

By Target Market

By Specialty

By Payment


📝 Reviews

Deep-dive reviews and benchmark reports:


⚠️ Blacklist

Gateways that have been flagged for serious issues (rug-pulls, data leaks, fake models).

See blacklist.md for the current list and evidence archive.

If you've been affected, please open a GitHub Discussion with evidence. We review reports transparently.


🧪 Benchmark

We run periodic benchmarks testing:

  • Latency (TTFT, total response time)
  • Throughput (tokens/sec)
  • Reliability (uptime %)
  • Accuracy (model behavior matches official?)
  • Pricing consistency (quoted vs. billed)

See scripts/benchmark.py for the test harness. Results in data/benchmarks/.


🤝 Contributing

We welcome PRs from users, gateway operators, and researchers. See CONTRIBUTING.md.

Quick ways to help:

  • ✏️ Update a gateway's pricing or model list
  • 📝 Submit a new gateway (use the template)
  • 🚨 Report a scam or outage
  • 🧪 Submit benchmark data
  • 🌐 Improve translations

📜 License

Content is released under the MIT License. Attribution appreciated but not required.


🙏 Acknowledgments

Inspired by apicompare.best, helpaio.com/transit, and trustmrr.com — each taught us one piece of the puzzle: price data, content depth, and verified trust. We combine them with GitHub-native openness.


Built by Mack Ding and contributors · Made for the global OPC community


🦞 OPC Ecosystem

Built by @MackDing — One-Person Company infrastructure powered by AI agents.

Project What it does
opc.ren OPC founder hub — tools, signals, community
CodexClaw Telegram bot for remote Codex access with MCP + subagent routing
awesome-ai-api Leaderboard of 200+ AI API gateways & relays
claude-context-health Diagnose & fix Claude Code session degradation
opc-daily-signal AI-powered daily decision intelligence for OPC founders
doc-preprocess-hub Enterprise document preprocessing — MinerU + docling

About

Curated list of 200+ AI API providers with daily ranking, OpenAI-compatible API checks, and live /v1/models probing for Claude, DeepSeek, Gemini, and more.

Topics

Resources

License

Contributing

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors

Languages