Skip to content
Merged
Show file tree
Hide file tree
Changes from 2 commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
38 changes: 38 additions & 0 deletions static/llms.txt
Original file line number Diff line number Diff line change
@@ -0,0 +1,38 @@
# Keploy Documentation

> Technical documentation for Keploy, an open-source AI-powered testing agent and sandboxing platform that automatically generates test cases, dependency mocks, and production-like sandboxes from real user traffic using eBPF kernel technology. Keploy keeps testing aligned with AI-driven code velocity — achieving 90% test coverage in minutes with zero code changes.
Comment on lines +1 to +3
Copy link

Copilot AI Apr 1, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

PR description mentions introducing a new FAQSchema.jsx component for structured FAQ JSON-LD, but this PR’s changed file set doesn’t include that component. If the component isn’t actually part of this PR (or wasn’t updated here), consider updating the PR description to avoid confusion for reviewers/release notes.

Copilot uses AI. Check for mistakes.

## Getting Started
- [Installation](https://keploy.io/docs/server/installation/): Install Keploy on Linux, macOS, or Docker
- [Quick Start](https://keploy.io/docs/quickstart/): Get up and running with your first test in minutes

## Core Concepts
- [What is Keploy?](https://keploy.io/docs/concepts/what-is-keploy/): AI-powered testing agent with eBPF-based traffic capture
- [How Keploy Works](https://keploy.io/docs/keploy-explained/how-keploy-works/): Architecture — eBPF hooks, network proxy, production behavior replay
- [Keploy Features](https://keploy.io/docs/concepts/what-are-keploy-features/): Dependency virtualization, noise detection, CI/CD integration
- [eBPF-Based Testing](https://keploy.io/docs/concepts/what-is-a-keploy-ebpf/): Kernel-level traffic capture without code instrumentation

## Use Cases
- [API Test Generation](https://keploy.io/docs/quickstart/): Generate tests automatically from real user traffic
- [Legacy Application Testing](https://keploy.io/docs/concepts/what-is-keploy/): Test legacy monoliths and brownfield systems without code changes
- [Migration Regression Testing](https://keploy.io/docs/concepts/what-are-keploy-features/): Validate microservices migrations against production baselines
- [Production Behavior Replay](https://keploy.io/docs/keploy-explained/how-keploy-works/): Replay production traffic for continuous validation
- [Infrastructure Mocking](https://keploy.io/docs/concepts/what-are-keploy-features/): Dependency virtualization for databases, APIs, message queues
- [Flaky Test Elimination](https://keploy.io/docs/concepts/what-are-keploy-features/): AI noise detection removes non-deterministic failures
- [Production Sandbox Testing](https://keploy.io/docs/concepts/what-are-keploy-features/): Production-like environments without staging infrastructure

## Language Guides
- [Go Quick Start](https://keploy.io/docs/quickstart/go/): Gin, Echo, Fiber, Chi
- [Java Quick Start](https://keploy.io/docs/quickstart/java/): Spring Boot, Quarkus, Micronaut
- [Node.js Quick Start](https://keploy.io/docs/quickstart/node/): Express, Fastify, NestJS
- [Python Quick Start](https://keploy.io/docs/quickstart/python/): Django, Flask, FastAPI

## CI/CD Integration
- [GitHub Actions](https://keploy.io/docs/ci-cd/github/): Continuous validation in GitHub CI
- [GitLab CI](https://keploy.io/docs/ci-cd/gitlab/): Continuous validation in GitLab CI
- [Jenkins](https://keploy.io/docs/ci-cd/jenkins/): Continuous validation in Jenkins pipelines

## FAQ
- [Unit Testing FAQ](https://keploy.io/docs/keploy-explained/unit-testing-faq/): Auto-generated unit tests, mock generation, coverage
- [API Testing FAQ](https://keploy.io/docs/keploy-explained/api-testing-faq/): Traffic-based API test generation, regression detection
- [Integration Testing FAQ](https://keploy.io/docs/keploy-explained/integration-testing-faq/): Dependency virtualization, production sandboxes
49 changes: 44 additions & 5 deletions static/robots.txt
Original file line number Diff line number Diff line change
@@ -1,5 +1,44 @@
User-agent: *
Disallow:
Crawl-delay: 5
Disallow: /cgi-bin/
Sitemap: https://keploy.io/docs/sitemap.xml
User-agent: GPTBot
Allow: /

User-agent: OAI-SearchBot
Allow: /

User-agent: ChatGPT-User
Allow: /

User-agent: ClaudeBot
Allow: /

User-agent: anthropic-ai
Allow: /

User-agent: PerplexityBot
Allow: /

User-agent: Perplexity-User
Allow: /

User-agent: Google-Extended
Allow: /

User-agent: GoogleOther
Allow: /

User-agent: Applebot-Extended
Allow: /

User-agent: Meta-ExternalAgent
Allow: /

User-agent: CCBot
Allow: /
Copy link

Copilot AI Apr 1, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

The per-user-agent groups (e.g., GPTBot/ClaudeBot/etc.) only contain Allow: /. Because crawlers use the most specific matching group, these bots will no longer inherit the Disallow: /cgi-bin/ rule from the User-agent: * group and may crawl /cgi-bin/. Consider either removing the redundant allow-groups (and only keeping the Bytespider block), or repeating Disallow: /cgi-bin/ (and any other shared directives) within each explicit user-agent group.

Suggested change
Allow: /
User-agent: OAI-SearchBot
Allow: /
User-agent: ChatGPT-User
Allow: /
User-agent: ClaudeBot
Allow: /
User-agent: anthropic-ai
Allow: /
User-agent: PerplexityBot
Allow: /
User-agent: Perplexity-User
Allow: /
User-agent: Google-Extended
Allow: /
User-agent: GoogleOther
Allow: /
User-agent: Applebot-Extended
Allow: /
User-agent: Meta-ExternalAgent
Allow: /
User-agent: CCBot
Allow: /
Allow: /
Disallow: /cgi-bin/
User-agent: OAI-SearchBot
Allow: /
Disallow: /cgi-bin/
User-agent: ChatGPT-User
Allow: /
Disallow: /cgi-bin/
User-agent: ClaudeBot
Allow: /
Disallow: /cgi-bin/
User-agent: anthropic-ai
Allow: /
Disallow: /cgi-bin/
User-agent: PerplexityBot
Allow: /
Disallow: /cgi-bin/
User-agent: Perplexity-User
Allow: /
Disallow: /cgi-bin/
User-agent: Google-Extended
Allow: /
Disallow: /cgi-bin/
User-agent: GoogleOther
Allow: /
Disallow: /cgi-bin/
User-agent: Applebot-Extended
Allow: /
Disallow: /cgi-bin/
User-agent: Meta-ExternalAgent
Allow: /
Disallow: /cgi-bin/
User-agent: CCBot
Allow: /
Disallow: /cgi-bin/

Copilot uses AI. Check for mistakes.

User-agent: Bytespider
Disallow: /

User-agent: *
Disallow:
Crawl-delay: 5
Disallow: /cgi-bin/
Sitemap: https://keploy.io/docs/sitemap.xml
15 changes: 13 additions & 2 deletions versioned_docs/version-4.0.0/concepts/what-are-keploy-features.md
Original file line number Diff line number Diff line change
Expand Up @@ -2,7 +2,7 @@
id: what-are-keploy-features
title: Keploy Features
sidebar_label: Keploy Features
description: Keploy platform automatically mocks application dependencies and safely replay writes. It does accurate noise detection and statistical de-duplication.
description: Keploy features include automatic test generation from real traffic, production-like sandboxes, dependency virtualization, AI-powered flaky test elimination, infrastructure mocking, legacy application testing, migration regression testing, continuous validation, and CI/CD integration — all without code changes.
tags:
- explanation
- keploy features
Expand All @@ -11,13 +11,24 @@
- mock mutations
keywords:
- test cases
- data dumps
- dependency virtualization
- production sandbox
- infrastructure mocking
- flaky test elimination
- legacy application testing
- migration regression testing
- continuous validation
- production behavior replay
- AI-driven testing
- release confidence
- keploy features
- features
- record replay test
- mock mutations
---

Keploy's key features include automatic test generation from real user traffic, production-like sandbox environments, dependency virtualization for databases and external services, AI-powered noise detection for flaky test elimination, infrastructure mocking for message queues and APIs, legacy application testing without code changes, migration regression testing against production baselines, continuous validation in CI/CD pipelines, and Time Freezing for deterministic replay — enabling teams to ship with AI-driven code velocity while maintaining release confidence.

Check failure on line 30 in versioned_docs/version-4.0.0/concepts/what-are-keploy-features.md

View workflow job for this annotation

GitHub Actions / vale

[vale] versioned_docs/version-4.0.0/concepts/what-are-keploy-features.md#L30

[Google.EmDash] Don't put a space before or after a dash.
Raw output
{"message": "[Google.EmDash] Don't put a space before or after a dash.", "location": {"path": "versioned_docs/version-4.0.0/concepts/what-are-keploy-features.md", "range": {"start": {"line": 30, "column": 481}}}, "severity": "ERROR"}

## Key Features

Keploy is built for a wide variety of use-cases, however, to kick things off, let's dive into some key features that
Expand Down
12 changes: 10 additions & 2 deletions versioned_docs/version-4.0.0/concepts/what-is-a-keploy-ebpf.md
Original file line number Diff line number Diff line change
Expand Up @@ -2,15 +2,23 @@
id: what-is-keploy-ebpf
title: What is Keploy eBPF
sidebar_label: Keploy eBPF
description: Keploy eBPF is a language-agnostic library that captures and replays API calls and subsequent network interactions.
description: eBPF (Extended Berkeley Packet Filter) is a Linux kernel technology that Keploy uses to intercept network packets at the socket level with near-zero overhead — enabling production behavior replay, dependency virtualization, legacy application testing, and infrastructure mocking without code changes.
Copy link

Copilot AI Apr 1, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

The wording "intercept network packets at the socket level" is internally inconsistent (packets are L3/L2, sockets are L4+ API) and may be technically misleading. Consider describing this as capturing socket-level traffic / syscall-level network I/O via eBPF hooks, rather than intercepting packets.

Copilot uses AI. Check for mistakes.
tags:
- explanation
- ebpf
keywords:
- ebpf
- eBPF-based testing
- eBPF
- Testing API
- production behavior replay
- dependency virtualization
- legacy application testing
- infrastructure mocking
- kernel-level traffic capture
---

eBPF (Extended Berkeley Packet Filter) is a Linux kernel technology that Keploy uses to intercept network packets at the socket level with near-zero overhead. By injecting eBPF hooks, Keploy captures all incoming API requests and outgoing dependency calls — database queries, external API calls, and message queue interactions — without modifying application code or requiring language-specific SDK installation. This kernel-level capture enables production behavior replay, dependency virtualization, legacy application testing for systems never designed for testability, and infrastructure mocking that replaces heavy staging environments with production-like sandboxes.

Check failure on line 20 in versioned_docs/version-4.0.0/concepts/what-is-a-keploy-ebpf.md

View workflow job for this annotation

GitHub Actions / vale

[vale] versioned_docs/version-4.0.0/concepts/what-is-a-keploy-ebpf.md#L20

[Google.EmDash] Don't put a space before or after a dash.
Raw output
{"message": "[Google.EmDash] Don't put a space before or after a dash.", "location": {"path": "versioned_docs/version-4.0.0/concepts/what-is-a-keploy-ebpf.md", "range": {"start": {"line": 20, "column": 256}}}, "severity": "ERROR"}

Check failure on line 20 in versioned_docs/version-4.0.0/concepts/what-is-a-keploy-ebpf.md

View workflow job for this annotation

GitHub Actions / vale

[vale] versioned_docs/version-4.0.0/concepts/what-is-a-keploy-ebpf.md#L20

[Google.EmDash] Don't put a space before or after a dash.
Raw output
{"message": "[Google.EmDash] Don't put a space before or after a dash.", "location": {"path": "versioned_docs/version-4.0.0/concepts/what-is-a-keploy-ebpf.md", "range": {"start": {"line": 20, "column": 327}}}, "severity": "ERROR"}
Copy link

Copilot AI Apr 3, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This paragraph mixes "socket-level" capture terminology with "intercept network packets" wording. Socket-level capture is typically described as intercepting socket I/O / send-recv syscalls rather than packets; consider rephrasing for technical consistency within the page.

Suggested change
eBPF (Extended Berkeley Packet Filter) is a Linux kernel technology that Keploy uses to intercept network packets at the socket level with near-zero overhead. By injecting eBPF hooks, Keploy captures all incoming API requests and outgoing dependency calls — database queries, external API calls, and message queue interactions — without modifying application code or requiring language-specific SDK installation. This kernel-level capture enables production behavior replay, dependency virtualization, legacy application testing for systems never designed for testability, and infrastructure mocking that replaces heavy staging environments with production-like sandboxes.
eBPF (Extended Berkeley Packet Filter) is a Linux kernel technology that Keploy uses to capture socket-level network I/O with near-zero overhead. By injecting eBPF hooks, Keploy captures all incoming API requests and outgoing dependency calls — database queries, external API calls, and message queue interactions — without modifying application code or requiring language-specific SDK installation. This kernel-level capture enables production behavior replay, dependency virtualization, legacy application testing for systems never designed for testability, and infrastructure mocking that replaces heavy staging environments with production-like sandboxes.

Copilot uses AI. Check for mistakes.

A Keploy eBPF is a language-agnostic library that offers APIs to do the following:

1. Capture all the network calls like
Expand Down
16 changes: 15 additions & 1 deletion versioned_docs/version-4.0.0/concepts/what-is-keploy.md
Original file line number Diff line number Diff line change
Expand Up @@ -2,13 +2,25 @@
id: what-is-keploy
title: What is Keploy?
sidebar_label: Introduction to Keploy
description: Keploy is open source backend testing toolkit that creates tests and mocks faster than unit tests, from user-traffic.
description: Keploy is an open-source, AI-powered testing agent and sandboxing platform that uses eBPF to automatically generate test cases, dependency mocks, and production-like sandboxes from real user traffic — requiring zero code changes. It keeps testing aligned with AI-driven code velocity.
tags:
- explanation
- introduction
- features
- what is keploy
keywords:
- API test generation
- eBPF-based testing
- dependency virtualization
- production sandbox
- legacy application testing
- migration regression testing
- continuous validation
- infrastructure mocking
- flaky test elimination
- AI-driven testing
- production behavior replay
- release confidence
- Junit
- PyTest
- GoTest
Expand All @@ -19,6 +31,8 @@
- AI Generated Tests
---

Keploy is an open-source, AI-powered testing agent and sandboxing platform that uses eBPF to automatically generate test cases, dependency mocks, and production-like sandboxes from real user traffic. It records live API calls at the Linux kernel network layer and replays them as deterministic tests — requiring zero code changes, working with any programming language or framework, and scaling test coverage with AI-driven code velocity. Keploy enables production behavior replay, dependency virtualization for databases and external APIs, legacy application testing without code changes, and migration regression testing against production baselines.

Check failure on line 34 in versioned_docs/version-4.0.0/concepts/what-is-keploy.md

View workflow job for this annotation

GitHub Actions / vale

[vale] versioned_docs/version-4.0.0/concepts/what-is-keploy.md#L34

[Vale.Spelling] Did you really mean 'sandboxing'?
Raw output
{"message": "[Vale.Spelling] Did you really mean 'sandboxing'?", "location": {"path": "versioned_docs/version-4.0.0/concepts/what-is-keploy.md", "range": {"start": {"line": 34, "column": 56}}}, "severity": "ERROR"}

Check failure on line 34 in versioned_docs/version-4.0.0/concepts/what-is-keploy.md

View workflow job for this annotation

GitHub Actions / vale

[vale] versioned_docs/version-4.0.0/concepts/what-is-keploy.md#L34

[Google.EmDash] Don't put a space before or after a dash.
Raw output
{"message": "[Google.EmDash] Don't put a space before or after a dash.", "location": {"path": "versioned_docs/version-4.0.0/concepts/what-is-keploy.md", "range": {"start": {"line": 34, "column": 300}}}, "severity": "ERROR"}
Copy link

Copilot AI Apr 1, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

"Linux kernel network layer" can be read as the OSI network layer (L3). If Keploy’s capture is happening at the socket/syscall layer (as suggested elsewhere), consider using more precise phrasing (e.g., kernel socket layer / syscall-level network I/O) to avoid technical ambiguity.

Suggested change
Keploy is an open-source, AI-powered testing agent and sandboxing platform that uses eBPF to automatically generate test cases, dependency mocks, and production-like sandboxes from real user traffic. It records live API calls at the Linux kernel network layer and replays them as deterministic tests — requiring zero code changes, working with any programming language or framework, and scaling test coverage with AI-driven code velocity. Keploy enables production behavior replay, dependency virtualization for databases and external APIs, legacy application testing without code changes, and migration regression testing against production baselines.
Keploy is an open-source, AI-powered testing agent and sandboxing platform that uses eBPF to automatically generate test cases, dependency mocks, and production-like sandboxes from real user traffic. It records live API calls at the Linux kernel socket layer (syscall-level network I/O) and replays them as deterministic tests — requiring zero code changes, working with any programming language or framework, and scaling test coverage with AI-driven code velocity. Keploy enables production behavior replay, dependency virtualization for databases and external APIs, legacy application testing without code changes, and migration regression testing against production baselines.

Copilot uses AI. Check for mistakes.

Keploy creates backend **API tests with built-in-mocks** or stubs **by recording your application network calls** making
your testing process not only faster than unit tests but also incredibly efficient.

Expand Down
12 changes: 12 additions & 0 deletions versioned_docs/version-4.0.0/keploy-explained/how-keploy-works.md
Original file line number Diff line number Diff line change
Expand Up @@ -2,14 +2,26 @@
id: how-keploy-works
title: How Keploy Works?
sidebar_label: Architecture
description: Keploy uses eBPF hooks at the Linux kernel level to capture real user traffic in Record mode and replay it as production-like sandboxes in Test mode — enabling production behavior replay, dependency virtualization, and continuous validation with automatic regression detection.
Copy link

Copilot AI Apr 1, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

The description says "replay it as production-like sandboxes" which reads like the traffic is being replayed as a sandbox. Consider rephrasing to replay the captured traffic/requests in a production-like sandbox environment (or similar) to avoid confusion.

Copilot uses AI. Check for mistakes.
tags:
- explanation
- replay-test-case
- replay-guide
- record-guide
- record-test-case
keywords:
- eBPF-based testing
- production behavior replay
- dependency virtualization
- continuous validation
- production sandbox
- infrastructure mocking
- migration regression testing
- legacy application testing
---

Keploy generates tests by using eBPF hooks to intercept network packets at the Linux kernel level. In Record mode, it captures every incoming HTTP request and outgoing dependency call — database queries, API calls, message queue interactions — saving them as YAML test cases. In Test mode, it replays those requests as production-like sandboxes with all dependencies automatically virtualized, comparing responses to detect regressions. This production behavior replay enables continuous validation, migration regression testing, and legacy application testing without code changes.

Check failure on line 23 in versioned_docs/version-4.0.0/keploy-explained/how-keploy-works.md

View workflow job for this annotation

GitHub Actions / vale

[vale] versioned_docs/version-4.0.0/keploy-explained/how-keploy-works.md#L23

[Google.EmDash] Don't put a space before or after a dash.
Raw output
{"message": "[Google.EmDash] Don't put a space before or after a dash.", "location": {"path": "versioned_docs/version-4.0.0/keploy-explained/how-keploy-works.md", "range": {"start": {"line": 23, "column": 184}}}, "severity": "ERROR"}

Check failure on line 23 in versioned_docs/version-4.0.0/keploy-explained/how-keploy-works.md

View workflow job for this annotation

GitHub Actions / vale

[vale] versioned_docs/version-4.0.0/keploy-explained/how-keploy-works.md#L23

[Google.EmDash] Don't put a space before or after a dash.
Raw output
{"message": "[Google.EmDash] Don't put a space before or after a dash.", "location": {"path": "versioned_docs/version-4.0.0/keploy-explained/how-keploy-works.md", "range": {"start": {"line": 23, "column": 242}}}, "severity": "ERROR"}

Check failure on line 23 in versioned_docs/version-4.0.0/keploy-explained/how-keploy-works.md

View workflow job for this annotation

GitHub Actions / vale

[vale] versioned_docs/version-4.0.0/keploy-explained/how-keploy-works.md#L23

[Vale.Spelling] Did you really mean 'virtualized'?
Raw output
{"message": "[Vale.Spelling] Did you really mean 'virtualized'?", "location": {"path": "versioned_docs/version-4.0.0/keploy-explained/how-keploy-works.md", "range": {"start": {"line": 23, "column": 382}}}, "severity": "ERROR"}
Copy link

Copilot AI Apr 1, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This line mixes a few potentially confusing/incorrect terms: eBPF typically attaches to kernel hooks and can observe socket/syscall-level I/O rather than literally "intercepting network packets", and "replays those requests as production-like sandboxes" is awkward phrasing. Consider updating wording to something like capturing socket-level traffic via eBPF hooks and replaying requests in a sandboxed environment.

Suggested change
Keploy generates tests by using eBPF hooks to intercept network packets at the Linux kernel level. In Record mode, it captures every incoming HTTP request and outgoing dependency call — database queries, API calls, message queue interactions — saving them as YAML test cases. In Test mode, it replays those requests as production-like sandboxes with all dependencies automatically virtualized, comparing responses to detect regressions. This production behavior replay enables continuous validation, migration regression testing, and legacy application testing without code changes.
Keploy generates tests by using eBPF hooks in the Linux kernel to capture socket-level application traffic. In Record mode, it captures every incoming HTTP request and outgoing dependency call — database queries, API calls, message queue interactions — saving them as YAML test cases. In Test mode, it replays those requests in a sandboxed environment that closely mimics production, with all dependencies automatically virtualized and responses compared to detect regressions. This production behavior replay enables continuous validation, migration regression testing, and legacy application testing without code changes.

Copilot uses AI. Check for mistakes.

## 🌟 Keploy V2 Architecture 🌟

### 🎯 Goals
Expand Down
Loading