|
Download README.md from OmniaKey/OmniaKey-AI-Coding-Tools: direct link, hf CLI and curl.
- Browser
- Download file 7.19 kB
-
https://huggingface.co/OmniaKey/OmniaKey-AI-Coding-Tools/resolve/main/README.md
- Command line
-
hf download hf://OmniaKey/OmniaKey-AI-Coding-Tools/README.md
-
curl -L -o README.md https://huggingface.co/OmniaKey/OmniaKey-AI-Coding-Tools/resolve/main/README.md
7.19 kB
| language: | |
| - zh | |
| - en | |
| license: mit | |
| tags: | |
| - tutorial | |
| - omniakey | |
| - ai-coding | |
| - claude-code | |
| - codex | |
| - cursor | |
| - cline | |
| - continue | |
| - aider | |
| - openai-compatible | |
| - anthropic | |
| - llm | |
| - vscode | |
| # OmniaKey AI Coding Tools | |
| > 在 Claude Code、Codex、Cursor、Cline、Continue、Aider 等 AI 编程工具中使用 Claude、GPT、Gemini、Grok 等模型。 | |
| [OmniaKey](https://omniakey.com/) 为 AI 编程工具提供一把统一 API Key,同时支持 Anthropic Messages、OpenAI Chat Completions 和 OpenAI Responses 等协议。按实际用量计费,无需为每个工具分别管理订阅。 | |
| > This repository is a configuration guide, not a downloadable model checkpoint. English quick start is included below. | |
| ## 快速入口 | |
| - [创建 OmniaKey API Key](https://omniakey.com/dashboard/tokens) | |
| - [查看模型目录](https://omniakey.com/models) | |
| - [AI 编程工具完整指南](https://docs.omniakey.com/cn/coding) | |
| - OpenAI-compatible Base URL:`https://api.omniakey.com/v1` | |
| - Anthropic Base URL:`https://api.omniakey.com` | |
| 请勿把真实 API Key 提交到 Git、截图、Issue 或日志中。示例统一使用 `sk-your-omniakey-key`。 | |
| ## 先检查当前可用模型 | |
| 模型和权限会持续更新。不要依赖旧教程中的静态价格表,先用自己的 Key 查询: | |
| ```bash | |
| curl https://api.omniakey.com/v1/models \ | |
| -H "Authorization: Bearer sk-your-omniakey-key" | |
| ``` | |
| 从返回结果复制准确的 model ID,再填入对应工具。 | |
| ## Claude Code | |
| Claude Code 使用 Anthropic Messages API,因此 Base URL **不要带 `/v1`**: | |
| ```bash | |
| export ANTHROPIC_BASE_URL="https://api.omniakey.com" | |
| export ANTHROPIC_AUTH_TOKEN="sk-your-omniakey-key" | |
| claude | |
| ``` | |
| Claude Code 会自行请求 `/v1/messages`。如果把 Base URL 写成 `https://api.omniakey.com/v1`,最终会变成错误的 `/v1/v1/messages`。 | |
| 进入会话后可以用 `/model` 显式选择当前可用的 Claude 模型。OmniaKey 不会静默把你指定的模型替换成其他模型。 | |
| ## OpenAI Codex CLI | |
| Codex 自定义 provider 使用 Responses API。编辑 `~/.codex/config.toml`: | |
| ```toml | |
| model = "gpt-5.6-sol" | |
| model_provider = "omniakey" | |
| [model_providers.omniakey] | |
| name = "OmniaKey" | |
| base_url = "https://api.omniakey.com/v1" | |
| env_key = "OMNIAKEY_API_KEY" | |
| wire_api = "responses" | |
| ``` | |
| 然后设置 Key 并启动: | |
| ```bash | |
| export OMNIAKEY_API_KEY="sk-your-omniakey-key" | |
| codex | |
| ``` | |
| 示例 model ID 可能随目录更新;排障时先确认它出现在当前 Key 的 `/v1/models` 返回中。 | |
| ## Cursor | |
| 在 Cursor 的模型或自定义 OpenAI 设置中填写: | |
| | 配置项 | 值 | | |
| | --- | --- | | |
| | OpenAI API Key | `sk-your-omniakey-key` | | |
| | Override OpenAI Base URL | `https://api.omniakey.com/v1` | | |
| | Model | 从 `/v1/models` 复制一个当前可用 ID | | |
| Cursor 不同版本对自定义模型、Agent 模式和 Tab Completion 的支持可能不同。若某功能仍走 Cursor 自有服务,这是客户端能力边界,并不代表 API Key 无效。 | |
| ## Cline | |
| 在 Cline 设置中选择 **OpenAI Compatible**,然后填写: | |
| | 配置项 | 值 | | |
| | --- | --- | | |
| | Base URL | `https://api.omniakey.com/v1` | | |
| | API Key | `sk-your-omniakey-key` | | |
| | Model ID | 从 `/v1/models` 复制 | | |
| 保存后建议先用一个小任务测试 streaming 和工具调用,再开始大型代码修改。 | |
| ## Continue | |
| 在 Continue 中新增一个 OpenAI-compatible 模型: | |
| ```yaml | |
| name: OmniaKey | |
| version: 1.0.0 | |
| schema: v1 | |
| models: | |
| - name: OmniaKey Coding | |
| provider: openai | |
| model: YOUR_MODEL_ID | |
| apiBase: https://api.omniakey.com/v1 | |
| apiKey: sk-your-omniakey-key | |
| ``` | |
| Continue 的配置格式会随版本变化;如果你的版本使用 UI 配置,填入相同的 Base URL、Key 和 Model ID 即可。 | |
| ## Aider | |
| 先设置 OpenAI-compatible 凭证和端点: | |
| ```bash | |
| export OPENAI_API_KEY="sk-your-omniakey-key" | |
| export OPENAI_API_BASE="https://api.omniakey.com/v1" | |
| aider --model openai/YOUR_MODEL_ID | |
| ``` | |
| Aider 通过 LiteLLM 识别 provider 前缀。若模型需要不同前缀或参数,请以当前 Aider/LiteLLM 文档和 `/v1/models` 返回为准。 | |
| ## OpenAI SDK | |
| ### Python | |
| ```python | |
| from openai import OpenAI | |
| client = OpenAI( | |
| base_url="https://api.omniakey.com/v1", | |
| api_key="sk-your-omniakey-key", | |
| ) | |
| response = client.chat.completions.create( | |
| model="YOUR_MODEL_ID", | |
| messages=[ | |
| {"role": "user", "content": "Review this function for correctness."} | |
| ], | |
| ) | |
| print(response.choices[0].message.content) | |
| ``` | |
| ### Node.js | |
| ```javascript | |
| import OpenAI from "openai"; | |
| const client = new OpenAI({ | |
| baseURL: "https://api.omniakey.com/v1", | |
| apiKey: "sk-your-omniakey-key", | |
| }); | |
| const response = await client.chat.completions.create({ | |
| model: "YOUR_MODEL_ID", | |
| messages: [ | |
| { role: "user", content: "Write tests for this TypeScript module." }, | |
| ], | |
| }); | |
| console.log(response.choices[0].message.content); | |
| ``` | |
| ## 常见问题 | |
| | 现象 | 检查 | | |
| | --- | --- | | |
| | `401 Unauthorized` | Key 是否正确、是否带 `Bearer `、是否复制了多余空格 | | |
| | `404 Not Found` | OpenAI-compatible 工具是否使用 `/v1`;Claude Code 是否错误地多加了 `/v1` | | |
| | Model not found | 用当前 Key 请求 `/v1/models`,复制准确 ID | | |
| | 工具调用失败 | 先确认客户端使用的协议和该模型的工具调用能力 | | |
| | Streaming 中断 | 检查本地代理、企业网络和客户端超时设置;保留 request ID 便于排查 | | |
| | `429` | 按响应中的 retry/reset 信息等待后重试,避免无上限自动循环 | | |
| ## 为什么选择 OmniaKey | |
| - 一把 Key 连接 Claude、GPT、Gemini、Grok 等模型 | |
| - 同时提供 Anthropic 原生协议与 OpenAI-compatible 接口 | |
| - 支持 Chat Completions、Responses、streaming 和工具调用 | |
| - 模型 ID 显式可见,不静默替换 | |
| - 一个 Dashboard 管理 Key、余额和用量 | |
| --- | |
| ## English quick start | |
| Create a key at [OmniaKey Dashboard](https://omniakey.com/dashboard/tokens), then choose the protocol your coding tool actually uses. | |
| ### Claude Code | |
| ```bash | |
| export ANTHROPIC_BASE_URL="https://api.omniakey.com" | |
| export ANTHROPIC_AUTH_TOKEN="sk-your-omniakey-key" | |
| claude | |
| ``` | |
| Do not append `/v1`; Claude Code appends `/v1/messages` itself. | |
| ### OpenAI-compatible tools | |
| Use these values in Cursor, Cline, Continue, Aider, SDKs, and other compatible clients: | |
| ```text | |
| Base URL: https://api.omniakey.com/v1 | |
| API Key: sk-your-omniakey-key | |
| Model: choose an ID returned by GET /v1/models | |
| ``` | |
| ### Codex | |
| ```toml | |
| model = "gpt-5.6-sol" | |
| model_provider = "omniakey" | |
| [model_providers.omniakey] | |
| name = "OmniaKey" | |
| base_url = "https://api.omniakey.com/v1" | |
| env_key = "OMNIAKEY_API_KEY" | |
| wire_api = "responses" | |
| ``` | |
| Always verify the current model catalog before relying on an example ID: | |
| ```bash | |
| curl https://api.omniakey.com/v1/models \ | |
| -H "Authorization: Bearer sk-your-omniakey-key" | |
| ``` | |
| ## Links | |
| - [OmniaKey](https://omniakey.com/) | |
| - [Get an API Key](https://omniakey.com/dashboard/tokens) | |
| - [Model catalog](https://omniakey.com/models) | |
| - [Coding tools guide](https://docs.omniakey.com/en/coding) | |
| This repository contains documentation only and is not affiliated with Anthropic, OpenAI, Cursor, Cline, Continue, or Aider. Product names belong to their respective owners. | |