Supported Site List
All API Hub primarily supports two types of objects: the intermediate stations/management panels you use daily, and backend systems you build yourself that you wish to continue managing within the plugin.
Sites That Can Be Added and Managed Daily
The following are the most common and suitable types of sites for unified management through All API Hub, spanning open-source distribution architectures, commercial relay services, and independent platforms.
Compatible Account Site Architectures
Sites built on mainstream open-source LLM API management and distribution systems, including their forks and derivative implementations:
| Site / System | Site Overview | Official Link |
|---|---|---|
| New API | Unified AI model aggregation and distribution center. | Official Website / GitHub |
| Sub2API | Sub2API-CRS2 one-stop open-source relay service, providing unified access to Claude, OpenAI, Gemini, and Antigravity subscriptions, with shared plans, more efficient cost sharing, and seamless use of native tools. | GitHub |
| One API | LLM API management and distribution system, supporting mainstream models such as OpenAI, Azure, Anthropic Claude, Google Gemini, DeepSeek, etc. Unified API adaptation, can be used for Key management and secondary distribution. | GitHub |
| one-hub | OpenAI interface management and distribution system, modified from songquanpeng/one-api, supports more models, adds a statistics page, and improves function calls for non-OpenAI models. | Official Website / GitHub |
| Veloera | This project has been discontinued. | GitHub |
| VoAPI / VoAPI v2 | The new VoAPI is a modern AI model management and distribution system based on the original VoAPI, while maintaining legacy deployment compatibility. | GitHub |
| DoneHub | This project is a secondary development based on one-hub. | GitHub |
| Super-API | Super-Api brand new AI model interface management and distribution system based on NewAPI. | GitHub |
| APIyi | A commercial AI model relay service platform based on the New API architecture. | — |
| ModelFlare | A high-performance AI model relay platform derived from New API. | — |
| v-api | A powerful intermediate platform based on the secondary development of one-api. | No data available |
Independent Platforms and Commercial Relays
AI model platforms and relay services with independent account systems, dedicated management consoles, or custom APIs:
| Site / System | Site Overview | Official Link |
|---|---|---|
| OpenRouter | An independent AI model aggregation platform offering a unified API for multiple model providers. | Official Website / Documentation / Setup Guide |
| Kimi Open Platform | The official Kimi (Moonshot) open platform; its international and China deployments have separate account systems. | International / China |
| AnyRouter | Claude Code intermediate station · Zero threshold · Free $50 | Documentation / Official Website |
| AgentRouter | AI API routing and relay service platform. | Official Website |
| AIHubMix | An independent AI API aggregation platform with model APIs and a dedicated account console. | Official Website / API Documentation / Setup Guide |
| LaoZhang API | A commercial AI model API relay with default, domestic/backup, direct international and CF-CDN endpoints sharing one account system. | — |
| Dahl | An AI inference service using token credits, with model APIs and an account console. | — |
| FoxCode | An AI coding model relay and subscription service platform for developers. | — |
| Xiaoen AI | An independent CharityAPI-based AI service platform with model APIs and an account console. | — |
| Cubence | A commercial AI model API platform with model access, groups and credit management. | — |
| RightCode | A commercial AI agent relay service; right.codes and rightapi.ai are entry points to the same service. | — |
| Grsai | A commercial AI model aggregation relay; grsai.com and grsai.ai are entry points to the same service. | — |
| FreeModel | An independent AI service platform offering model APIs, stored credits and subscription plans. | — |
| SharedChat | An AI subscription sharing service platform focused on pooled access and quota management for Codex and Vibe Code. | — |
| WONG Public Welfare Site | No stable public official description available. | No stable public official link available |
| Krill AI | An independent commercial AI model relay platform with model APIs and an account console. | — |
| EMRank | An independent gateway for embedding and reranker models with model APIs, an account console and daily check-in. | Site / Docs |
| RelayFor | An independent AI public welfare service platform offering model APIs and an account console. | Official Website |
| Daya API | A unified AI API access platform aggregating multiple model providers, with model APIs, standalone keys, and credit management. | — |
| PM-API | A community AI relay with model APIs, token key management, redemption codes, and daily check-in; pmcat.top and pap.pmcat.top are two deployments of one program. | Site |
| Requesty | A hosted LLM gateway with unified model routing, prepaid credits and key management. | Site / Docs |
| Vercel AI Gateway | Vercel's official AI model gateway, calling many model providers with one API key, with credits, request logs and budgets. | Website / Docs |
❤️ Recommended Sponsored Sites
If you are looking for stable, efficient, and highly compatible AI relay services, try our partners:
Qiniu Cloud AI is Qiniu Cloud's enterprise MaaS platform, offering one-stop access to 150+ mainstream models worldwide with compatibility across major provider protocols and full-modal processing for text, image, audio, video, and files. Enterprise users can claim 12 million free tokens via this link, with referral rewards up to tens of billions of tokens.
FennoAI is a stable and efficient API relay provider focused on Codex relay services. It supports the OpenAI and Anthropic protocols, integrates with popular coding tools such as Codex, Claude Code, and OpenCode, reliably handles enterprise demand at the scale of 100 billion tokens per day, and supports business-to-business settlement and invoicing for entities in China and overseas. Subscribe through the dedicated link for just $1.99 to receive $50 in Coding Plan credits. Referral purchases earn up to 20% commission, with higher rewards as you invite more people (setup guide).
PackyCode offers relay services for Claude Code, Codex, Gemini, and more. Register using this link and enter the "all-api-hub" promo code during first recharge to get 10% off (setup guide).
Xingchen AI is a stable and efficient API relay provider offering services for Claude Code, Codex, Gemini, and more. It supports 1:1 top-ups, provides invoices, and offers Claude at as low as 40% of the standard price. Learn more and start using it through this link (setup guide).
AICodeMirror provides official high-stability relay services for Claude Code / Codex / Gemini CLI, with enterprise-grade concurrency, fast invoicing, and 24/7 dedicated technical support. Claude Code / Codex / Gemini official channels are available from 38% / 2% / 9% of the original price, with extra discounts on top-ups. Register via this link to enjoy 20% off your first top-up, and enterprise customers can get up to 25% off.
Suixiang AI Relay is a reliable and efficient API relay provider for Claude, Codex, Gemini, and more. It supports pay-as-you-go 1:1 top-ups, daily check-in test credits, redundant routes, cross-region disaster recovery, and automatic failover. Learn more through this link (setup guide).
APIMart is a low-cost API platform for AI image & video generation — GPT-Image-2 from $0.006/image, 160+ images per dollar. One async API covers both image and video: submit a task, get an ID, then fetch results via polling or callback. Batch tens of thousands of images without timeouts, switch models without changing code. Pay-as-you-go with no monthly fee. Sign up through this link to get started.
88API Token Aggregation Platform A one-stop multi-model API platform operated by a Hong Kong company for developers, creators, and AI application users. Its unified API connects text, image, speech, and video models for AI coding, translation, content creation, voiceovers, image generation, and video generation. Add and switch providers directly in CC Switch, with major international payment methods, invoicing, and stable enterprise services. Register through the dedicated link to receive bonus credits!
Dola Seed on BytePlus ModelArk is ByteDance's full-modal general large model for the global market, available through BytePlus ModelArk with multimodal understanding and generation, agent collaboration, reasoning, tool integration, and coding capabilities. Register through the BytePlus ModelArk link to get 500,000 free inference tokens per model.
Compatible Variants with Limited Public Information
The following compatible types are currently within the plugin's support scope, but have relatively limited public official information. They are typically found in private deployments, closed-source branches, or branded versions of specific sites:
Neo-APIRIX_API
If these systems do not have a stable public homepage, no additional links will be provided here. Their usability usually depends on whether the target site is a commonly compatible version.
Self-Built Backends You Can Connect and Continue Managing
If you are also building your own backend system, All API Hub supports importing the current site into your chosen backend for continued management.
| Backend System | Site Overview | Official Link |
|---|---|---|
| OmniRoute | Self-hosted AI gateway with a large built-in provider catalogue, scoped access tokens, and connection-level upstream address overrides. | GitHub |
| CLIProxyAPI | Encapsulates Gemini CLI, Antigravity, ChatGPT Codex, Claude Code, Qwen Code, and iFlow into API services compatible with OpenAI / Gemini / Claude / Codex. | Documentation / GitHub |
| New API | Unified AI model aggregation and distribution center. | Official Website / GitHub |
| Sub2API | Sub2API-CRS2 one-stop open-source relay service, providing unified access to Claude, OpenAI, Gemini, and Antigravity subscriptions, with support for shared plans and native tools. | GitHub |
| Magpie | A model gateway for AI coding tools, with menu-bar controls to switch models for Codex, Claude Code and other agents. | GitHub |
| gpt-load | Lightweight self-hosted AI gateway (Go); a group is one channel driver plus its credential pool. | GitHub |
| AxonHub | Open-source AI Gateway, callable via any SDK for 100+ LLMs, with built-in failover, load balancing, cost control, and full-link tracing. | Official Website / GitHub |
| Claude Code Hub | A multi-vendor AI API proxy and operating platform for teams, unifying access to Claude, OpenAI Compatible, Codex, and Gemini, with support for elastic scheduling, monitoring, and price management. | GitHub |
| Octopus | Personal LLM API aggregation service. | GitHub |
| Veloera | This project has been discontinued. | GitHub |
| DoneHub | This project is a secondary development based on one-hub. | GitHub |
Batch importing multiple keys from one site
Choose “Create one new channel per key” or “Combine same-site keys into one new channel” during batch import. The choice is remembered per target management connection. The grouped preview shows source keys, resource count, and the union of model lists. Different model permissions produce a notice; users decide whether to combine keys, and the target gateway owns scheduling. Different inference endpoints (including paths) or native channel types remain separate.
| Managed site | Combined import scope |
|---|---|
| Magpie | Combined creation and appending missing keys to existing channels with the same address and protocol |
| New API, Octopus, gpt-load | Native multi-key resource creation |
| AxonHub | Regular API-key channel types |
| CLIProxyAPI | openai-compatibility resources |
| Sub2API, OmniRoute, Claude Code Hub, Veloera, DoneHub | Separate creation only |
Magpie also supports “Prefer adding to an existing channel”. A single channel with the same address and protocol is selected automatically; choose a target when several match, or choose to create a new one. Existing names, configuration, and model settings are preserved; only missing keys are added through native key actions. Without a matching channel, a combined channel is created. The target is read again before appending; an address or protocol change stops the operation. Check the target after partial or uncertain writes before importing again.