AI Model Aggregation · Built on New API
ByeCore 工作室是一个独立运营的 AI 模型聚合服务。我们把分散在各家厂商的模型能力，收敛到统一的入口、统一的鉴权与统一的计费之下——让你用一把密钥，从容调用整个模型世界。
ByeCore is an independently operated AI model-aggregation service. We bring the model capabilities scattered across providers under one endpoint, one authentication scheme and one billing model—so a single key gives you composed access to the entire model world.
使命 · MissionMission
把复杂的接入，收敛为一行配置；把分散的算力，聚成一个入口。
Reduce complex integration to a single line of configuration; gather scattered compute into one endpoint.
模型的世界正在迅速扩张：不同的厂商、不同的协议、不同的密钥与账单。开发者本不应为此疲于奔命——在十几个 SDK 之间切换，为每一家上游单独充值、单独对账、单独排障。
The world of models is expanding fast: different providers, different protocols, different keys and bills. Developers should not have to scramble through it all—switching between a dozen SDKs, topping up, reconciling and troubleshooting each upstream on its own.
ByeCore 把这一切抽象到一层之下。一个统一入口、一套鉴权、一份清晰的账单，背后是经过适配与路由的多家上游模型。我们要做的，是让接入这件事退回到它应有的位置——简单、稳定、可预期，然后把舞台留给你真正在构建的东西。
ByeCore abstracts all of it beneath a single layer. One unified endpoint, one authentication scheme, one clear bill—backed by multiple upstream models that have been adapted and routed. Our work is to return integration to where it belongs—simple, stable and predictable—and then leave the stage to what you are actually building.
我们做什么 · What we doWhat we do
我们既不训练模型，也不替模型背书；我们专注做好中间那一层——把请求接得稳、转得准、算得清。
We neither train models nor endorse them; we focus on the layer in between—accepting requests reliably, routing them accurately and accounting for them clearly.
所有调用经由单一域名 api.byecore.com 进出，统一鉴权、统一密钥，无需为每家上游分别维护接入。
Every call enters and exits through a single domain, api.byecore.com, with unified authentication and a single key—no need to maintain separate integrations for each upstream.
同时兼容 OpenAI 与 Anthropic 两套主流协议，既有的客户端与代码几乎零改动即可切换接入。
Compatible with both major protocols, OpenAI and Anthropic, so your existing clients and code can switch over with virtually no changes.
聚合 OpenAI、Claude、Gemini 等诸多厂商的模型能力，并在多条线路之间做路由、均衡与故障转移。
Aggregates the model capabilities of many providers—OpenAI, Claude, Gemini and more—with routing, balancing and failover across multiple lines.
用量、余额、密钥与账单尽在一处。预付费额度实时扣减，口径统一，账目清晰可查。
Usage, balance, keys and bills, all in one place. Prepaid credit is deducted in real time on a single, consistent basis, with clear and auditable records.
https://api.byecore.com
模型与上游 · Models & ProvidersModels & Providers
无论你习惯哪一套接口规范，都能用同一把密钥访问背后众多上游模型。
Whichever interface convention you prefer, the same key reaches the many upstream models behind it.
面向以 OpenAI SDK 与 /v1 规范构建的生态，Codex CLI、Cursor、Cline、Continue、Aider 等开箱即用。
For ecosystems built on the OpenAI SDK and the /v1 convention—Codex CLI, Cursor, Cline, Continue, Aider and more work out of the box.
面向 Anthropic Messages 规范，Claude Code、Claude Code Router 等可直接将入口指向本网关。
For the Anthropic Messages convention—Claude Code, Claude Code Router and others can point their endpoint straight at this gateway.
我们的信念 · PrinciplesPrinciples
本服务由 ByeCore 工作室独立运营，基于开源网关 New API 构建。我们恪守适用的法律法规，也恳请每一位用户在合法、合规的前提下使用——共同维护清朗、可信、可持续的环境。
This service is operated independently by ByeCore Studio and built on the open-source gateway New API. We adhere to applicable laws and regulations, and we ask every user to use it lawfully and in compliance—so that together we keep the environment clean, trustworthy and sustainable.
里程碑 · MilestonesMilestones
从首个版本到今天，我们持续打磨接入体验、计费口径与数据守护。
From the first release to today, we have kept refining the integration experience, our billing basis and our data protection.
ByeCore 工作室成立，首版《用户服务协议》发布，聚合接入服务正式对外提供。
ByeCore Studio was founded, the first Terms of Service was published, and the aggregation service became publicly available.
引入预付费额度与统一计费结算，完善数据处理与隐私引用，账目走向清晰统一。
We introduced prepaid credit and unified billing settlement, refined data handling and privacy references, and brought records toward clarity and consistency.
细化计费维度与退款规则，补充速率限制与公平使用策略，运营更稳健可预期。
We refined billing dimensions and refund rules and added rate limits and a fair-use policy, making operations steadier and more predictable.
服务协议重构为章·条结构，首部独立《隐私政策》发布——默认不用于训练、零数据保留可选。
The Terms of Service were reworked into a chapter-and-clause structure, and the first standalone Privacy Policy was published—not used for training by default, with zero data retention optional.
联系我们 · ContactContact
咨询、合作、答疑或行使你的数据权利，都欢迎通过以下方式找到我们。
For enquiries, partnerships, support or to exercise your data rights, you are welcome to reach us through the channels below.
目录Contents
01使命Mission 02我们做什么What we do 03模型与上游Models & providers 04信念Principles 05里程碑Milestones 06联系我们Contact