Keploy

io.github.keploy由 GitHub 账号 keploy 发布v1.0.0更新于 Nov 2, 2026

End-to-end API testing — generate and run tests from OpenAPI, curl, Postman, or real user traffic.

已验证Streamable HTTP可网页运行开发者工具安全与监控

概览

AI 生成的概览

Keploy 的远程 MCP 端点，可从 OpenAPI、curl、Postman 或真实流量生成并运行端到端 API 测试。

可以做什么
  • 根据 OpenAPI 规范、curl 命令、Postman 集合或真实用户流量生成 API 测试
  • 把录制的 API 与集成流程连同数据 mock 一起回放为测试
  • 借助 AI 扩展 API 覆盖：探测边界值、缺失字段、错误类型和乱序调用
  • 录制并回放数据库、消息队列和外部 API 交互，无需重新搭建基础设施即可运行测试
  • 汇总 API、集成、单元与端到端覆盖率报告
适用场景
适合希望让助手为现有服务搭建或扩充自动化 API 与集成测试套件的场景，尤其是已有接口规范或流量录制、想在不写测试代码的情况下覆盖 schema 与业务流程时。
运行要求
通过 streamable HTTP 访问的远程 MCP 端点，需要 OAuth 认证。需要 Keploy API 密钥，以 Bearer 令牌形式放在 Authorization 请求头中，密钥在 Keploy 账户设置中生成。需要能访问 Keploy 服务的网络。
安装前请注意
Authorization 请求头携带 Keploy API 密钥，应视为机密并避免出现在共享日志中；该密钥生成后仅显示一次。测试生成与覆盖率分析在 Keploy 托管服务上完成，因此 API schema 和来自流量的数据会发送给该第三方。

安装

在 SourceWeft 中

  1. 打开 控制台中的 Keploy，将其添加到工作区。
  2. 连接 所需的 凭据。它们会被私密存储，绝不会在公开页面上展示。
  3. 为需要使用其工具的对话启用该服务。

Web executable，通过 Streamable HTTP。 远程服务在工作区中配置后即可从网页运行时运行。

交给智能体安装

把这段提示词粘贴到 Claude Code、Codex、Cursor 等智能体，或 SourceWeft 对话中。智能体会读取该服务器的安装指南，先向你展示发布者、端点和登录方式，经你同意后再把它添加到你的客户端。密钥不会写进任何命令。

阅读 https://sourceweft.com/mcp/io.github.keploy%2Fmcp/install.md，并按其中说明把这个 MCP 服务器接入我的客户端。修改任何配置之前，先告诉我它的发布者、端点和登录方式，等我确认。不要把 API 密钥写进命令或我会提交的文件；问我密钥应从哪里获取。

查看智能体所依据的安装指南

其他 MCP 客户端

把它添加到你客户端的 mcpServers 配置中。 客户端会在首次连接时完成 OAuth 登录。

{
  "mcpServers": {
    "keploy": {
      "type": "http",
      "url": "https://api.keploy.io/client/v1/mcp"
    }
  }
}

README

[Keploy Logo]

[keploy%2Fkeploy | Trendshift]

⚡️ API tests faster than unit tests, from user traffic ⚡️

🌟 The must-have tool for developers in the AI-Gen era for 90% test coverage 🌟

[Slack] [LinkedIn] [YouTube] [X]

[Keploy CNCF Landscape] [GitHub Stars] [Help us reach 20k stars!]

Keploy is a developer‑centric API and integration testing tool that auto‑generates tests and data‑mocks faster than unit tests.

It records API calls, database queries, and streaming events — then replays them as tests. Under the hood, Keploy uses eBPF to capture traffic at the network layer, but for you it’s completely code‑less and language‑agnostic.

[Convert API calls to API tests test cases and Data Mocks using AI]

🐰 Fun fact: Keploy uses itself for testing! Check out our swanky coverage badge: [Coverage Status]  

Key Highlights

🎯 No code changes

Just run your app with keploy record. Real API + integration flows are automatically captured as tests and mocks. (Keploy uses eBPF under the hood to capture traffic, so you don’t need to add any SDKs or modify code.)

📹 Record and Replay complex Flows

Keploy can record and replay complex, distributed API flows as mocks and stubs. It's like having a very light-weight time machine for your tests—saving you tons of time!

👉 Read the docs on record-replay

[Convert API calls to test cases]

🐇 Complete Infra‑Virtualization (beyond HTTP mocks)

Unlike tools that only mock HTTP endpoints, Keploy records databases (Postgres, MySQL, MongoDB), streaming/queues (Kafka, RabbitMQ), external APIs, and more.

It replays them deterministically so you can run tests without re‑provisioning infra.

👉 Read the docs on infra virtualisation

[Convert API calls to test cases]

🧪 Combined Test Coverage

If you’re a developer, you probably care about statement and branch coverage — Keploy calculates that for you.

If you’re a QA, you focus more on API schema and business use‑case coverage — Keploy calculates that too. This way coverage isn’t subjective anymore.

👉 Read the docs on coverage

[ai test gen for api statement schema and branch coverage]

🤖 Expand API Coverage using AI

Keploy uses existing recordings, Swagger/OpenAPI Schema to find: boundary values, missing/extra fields, wrong types, out‑of‑order sequences, retries/timeouts.

This helps expand API Schema, Statement, and Branch Coverage.

👉 Read the docs on coverage

[ai test gen for api statement schema and branch coverage]

Other Capabilities

  • 🌐 CI/CD Integration: Run tests with mocks anywhere you like—locally on the CLI, in your CI pipeline (Jenkins, Github Actions..) , or even across a Kubernetes cluster. Read more

  • 🎭 Multi-Purpose Mocks: You can also use Keploy-generated Mocks, as server Tests!

  • 📊 Reporting: Unified reports for API, integration, unit, and e2e coverage with insights directly in your CI or PRs.

  • 🖥️ Console: A developer-friendly console to view, manage, and debug recorded tests and mocks.

  • ⏱️ Time Freezing: Deterministically replay tests by freezing system time during execution. Read more

  • 📚 Mock Registry: Centralized registry to manage, reuse, and version mocks across teams and environments. Read more

Quick Start

1. Install Keploy Agent

bash
source /dev/stdin <<<"$(curl -fsSL https://keploy.io/install.sh || echo 'echo Keploy installer download failed >&2; return 1')"

2. Record Test Cases

Start your app under Keploy to convert real API calls into tests and mocks.

bash
keploy record -c "CMD_TO_RUN_APP"

Example for Python:

bash
keploy record -c "python main.py"

3. Run Tests

Run tests offline without external dependencies.

bash
keploy test -c "CMD_TO_RUN_APP" --delay 10

Resources

- 📘 Installation

- 🏁 QuickStarts

Languages & Frameworks (Any stack)

Because Keploy intercepts at the network layer (eBPF), it works with any language, framework, or runtime—no SDK required.

Note: Some of the dependencies are not open-source by nature because their protocols and parsings are not open-sourced. It's not supported in Keploy enterprise.

[图片] [图片] [图片] [图片] [图片] [图片] [图片] [图片] [图片] [图片] [图片] [图片] [图片] [图片] [图片] [图片] [图片] [图片] [图片] [图片] [图片] [图片] [图片] [图片] [图片]

Questions?

Book a Live Demo / Enterprise Support

Want a guided walkthrough, dedicated support, or help planning enterprise rollout?

[图片]   [图片]

Prefer a calendar invite? Mention your availability in the email—we’ll send a calendar invite right away.

Documentation & Community

Contribute & Collaborate

Whether you're new or experienced, your input matters. Help us improve Keploy by contributing code, reporting issues, or sharing feedback.

Together, let's build better testing tools for modern applications.

来源：README.md，提交 558559c

工具

0
该服务器需要登录，因此 SourceWeft 无法在没有账号的情况下读取其工具，也未收录。已连接账号所能看到的内容可能因账号及其权限而异。

版本历史

1
  1. v1.0.0最新Sep 16, 2026

评分

已满足适用检查项的 85%

该评分衡量此收录条目的完整程度，依据注册中心条目、端点与公开来源计算得出。它不是对服务器本身的评价，也不表示服务器安全或当前可用：可连通性是带日期的检查结果，评分也不包含安装量或用户评价。

  • 近 30 天内可连通必需25 分
  • 已声明认证方式必需10 分
  • 已列出工具15 分

    请让服务器在无凭据时响应 tools/list，或发布可在沙箱中无凭据启动的软件包。

  • 描述必需15 分
  • 图标10 分
  • 文档15 分
  • 许可证10 分