docs: one Keploy across the site - no editions, accurate native macOS/Windows, eBPF on Linux - #918
Open
slayerjain wants to merge 1 commit into
Open
slayerjain wants to merge 1 commit into
slayerjain wants to merge 1 commit into
Conversation
…/Windows, eBPF on Linux Follow-up to #916, which fixed the install pages; this covers the rest of the current docs. - No edition names in user-facing text: CI/CD pages no longer demand an "Enterprise binary ... not the open-source" one (they install keploy.io/install.sh and sign in with KEPLOY_API_KEY); "Keploy Enterprise" as a product, "OSS" chips, "Start OSS Quickstart" and the Enterprise install reminder are gone. Plan availability is stated with the plan names (Free, Teams, Scale, Enterprise); ProductTier chips that said "Open Source" now list every plan, and "Team" is "Teams". - eBPF is the Linux mechanism, not how Keploy works everywhere: macOS and Windows intercept in user space. The intro/concept pages, the home page meta, the site schema and the llms.txt / llms-full.txt text (what AI assistants read) say so, and no longer claim Windows needs WSL2. - Quickstarts: the local-run sections are classified by what the app talks to - native on Linux, macOS (Apple Silicon) and Windows (x86-64) when that is HTTP/HTTPS, MySQL or MongoDB (with the macOS launcher and Windows cmd.exe caveats where the commands need them), Linux/WSL or Docker otherwise. - Fixed along the way: `keploy enterprise record` (no such command), GitHub release tarballs and keploy.io/ent/install.sh offered for install, the time-freezing page treating WSL as native, kernel 5.15 in llms text (the CLI requires 5.10), broken container names and sudo in Docker commands, dead quickstart anchors, and an Apache-2.0 license claim about the binary users install. Signed-off-by: slayerjain <shubhamkjain@outlook.com>
3 tasks
slayerjain
added a commit
that referenced
this pull request
Sep 23, 2026
* docs: describe the VS Code extension 3.0.0 as API and DB mocks The docs still sold the Keploy VS Code extension as a one-click unit test generator: running-keploy/utg-vscode-extension promised a "Generate Unit Tests" button, and integrations-vscode said the extension generates, views and manages AI unit tests. Extension 3.0.0 is a different product. It records the calls a test command makes to APIs and databases and replays them with those dependencies switched off, through `keploy mock`. It installs the Keploy CLI, works without an account, can set up AI coding agents, and writes CI jobs. It no longer generates unit tests, and no docs page described what it does. Rewrite the extension page from the extension's own behaviour: what it is for, requirements (VS Code 1.104+, Apple Silicon only on macOS, Linux kernel 5.10+ and glibc 2.34+ for the released CLI, Windows x64), the first-run CLI install, record and replay per platform, optional sign-in and what it adds, agent skill consent and where files land, CI setup, settings, the telemetry opt-out and what it does not cover yet, and the failure messages the panel shows. The page keeps its id, so the live URL keeps working; the S3 deploy has no working redirects. Point integrations-vscode at the new page, and stop the unit test generator pages (introduction, PR agent, FAQ, best practices, glossary) from sending readers to the extension for unit tests. The PR Agent is where unit test generation lives now. Signed-off-by: slayerjain <shubhamkjain@outlook.com> * docs: list the VS Code extension page under Integration Testing The rewritten extension page describes recording and replaying API and database mocks, but the sidebar still filed it under Unit Testing > Integration. Its breadcrumb therefore read "Unit Testing > Integration > VS Code Extension" directly above the note saying the extension no longer generates unit tests. Move it next to Mock Your Tests, the keploy mock page the extension drives, so the sidebar and breadcrumb match what the page describes. Signed-off-by: slayerjain <shubhamkjain@outlook.com> * docs: say which folders the agent skill question actually names The extension page claimed the "Just this project" button lists the repository folders it writes to. It does not: the button names only the folders of agents with no home-folder location (Cursor and GitHub Copilot), while the install it runs (`keploy skill install --project`) also writes .claude/skills or .agents/skills for every other agent it found. With only Claude Code or Codex detected the button carries no folder at all. A consent prompt must not be described as disclosing more than it does. State what the question names, what each answer writes, and point at the table that lists every repository folder. Signed-off-by: slayerjain <shubhamkjain@outlook.com> * docs: point Mock Your Tests readers at the VS Code extension guide Every docs link inside the VS Code extension opens Mock Your Tests: the Keploy: Mocking Docs command, the "What to check" button when a recording captured nothing, and the fallback of every other docs button. That page never mentioned the extension, so a user sent there from a failing panel found nothing about the panel or its messages. Add one line near the top that links the extension guide and its troubleshooting table. Signed-off-by: slayerjain <shubhamkjain@outlook.com> * docs: stop telling AI crawlers the VS Code extension generates tests llms-full.txt, which this repo builds for AI search engines, still said unit test generation is "available via VS Code extension, JetBrains plugins, CLI, and GitHub PR Agent", and listed the extension as "inline test generation and execution". Extension 3.0.0 records and replays API and database mocks and generates no tests. There is no Keploy JetBrains plugin on the JetBrains Marketplace, and neither Keploy CLI has a unit test generation command. Describe the extension as it is, link its guide, and name the PR Agent as the way to generate unit tests. content/llms-root.txt:58 carries the same claim but is left for a follow-up because it conflicts with the open #918. Signed-off-by: slayerjain <shubhamkjain@outlook.com> * docs: correct the signed-in registry note and the empty-recording fix Two statements on the VS Code extension page did not match the code: - "Depending on your plan" the CLI keeps mock sets in the Mock Registry. The enterprise CLI makes every signed-in, non --local `keploy mock` run registry-first with no plan check, and the api-server upload and download routes check the role and app, not the plan. What the page left out is that an upload or download failure falls back to the set on disk, and that signed-in replays are reported to the account as usage while --local runs are not. - "That run captured nothing" gave no macOS cause. On macOS the panel names npm, npx, yarn or pnpm when one of them started the tests, because macOS strips Keploy's instrumentation across that launcher. The panel prefills `npm test`, so this is the first failure a Mac Node user sees. Say what it means and what to run instead. Signed-off-by: slayerjain <shubhamkjain@outlook.com> * docs: say what a signed-in recording does to git and CI The VS Code extension page said an ignored recording stays on your machine, and that committing keploy/ lets teammates and CI replay it. For a signed-in user neither is true. The extension drops --local only when you are signed in, and then the enterprise CLI uploads the whole mocks.yaml, credentials included, to the Mock Registry whatever .gitignore says. After that upload, or after a signed-in download, it appends /*/mocks.yaml to keploy/.gitignore. That line hides every set git does not already track, so the CI job the extension writes, which checks for keploy/<set>/mocks.yaml and replays with --local, fails for lack of a recording. Scope the local-only statements to runs without an account. Say what a signed-in run uploads and adds to keploy/.gitignore, and that a tracked recording stays tracked. Give git add -f as the way to get such a set into CI, and mention the keploy.yml that the first run writes. Reproduced with a unit test against enterprise main's mockRegistry.upload: with keploy/ in the root .gitignore, the upload carried the credential line and keploy/.gitignore gained /*/mocks.yaml. A scratch repository then showed that git add keploy/ skipped the new set and kept staging the tracked one. Signed-off-by: slayerjain <shubhamkjain@outlook.com> * docs: keep the VS Code page true across pending extension fixes Several statements on the page describe extension behaviour that changes before 3.0.0 ships: - which folders the agent question and its "Just this project" label name, and that the fallback rules writer writes without a question - that the telemetry opt-out does not reach the CLI - the list of headlines a failed replay can show Keep only what holds before and after those changes. The page now says where each agent answer writes, which files the fallback writes, and how to turn off the CLI's own telemetry (disableTele in keploy.yml or KEPLOY_DISABLETELE). The CLI's sign-in events ignore both. The advice for runs that Keploy itself stopped moves to the "Keploy stopped that run." notification, which stays. The empty-recording row also names the macOS launcher trap. A runner started through npm, npx, yarn or pnpm, or through its own env-node shebang, loses Keploy's instrumentation, and node node_modules/.bin/jest keeps it. Signed-off-by: slayerjain <shubhamkjain@outlook.com> * docs: give the real size of the Keploy CLI download The page said the CLI is a single file of a few hundred MB. The released 3.8.44 builds are 160-191 MB: Linux x86-64 190.7 MB, Linux arm64 177.5 MB, macOS arm64 159.9 MB, and Windows x64 168.8 MB, going by the content-length of each download. Say about 200 MB. Signed-off-by: slayerjain <shubhamkjain@outlook.com> * docs: drop the CLI method from the unit test FAQ FAQ answers 10 and 11 offered a "CLI Tool" way to generate unit tests and said it needs an API key. Neither the open-source nor the enterprise Keploy CLI has a unit test generation command (the enterprise test-gen command generates API tests), and llms-full-root.txt on this branch already says unit tests come from the PR Agent. Point both answers at the PR Agent. Signed-off-by: slayerjain <shubhamkjain@outlook.com> * docs: say which telemetry opt-out reaches the Keploy agent The VS Code page told readers that KEPLOY_DISABLETELE=true turns off the Keploy CLI's usage reporting. On native Linux that is only half true. Unless the CLI runs as root, it starts its agent with sudo, and sudo's env_reset drops the variable, so the agent still reports usage. That is the default setup for the extension on Linux. disableTele: true in keploy.yml does reach that agent, because sudo keeps the working directory and the agent reads the same keploy.yml. For a Docker command the agent runs in its own container. It gets a fixed environment and no keploy.yml, so neither setting reaches it. Say which setting covers what, and keep the note that sign-in events follow neither. Signed-off-by: slayerjain <shubhamkjain@outlook.com> --------- Signed-off-by: slayerjain <shubhamkjain@outlook.com>
This branch has not been deployed
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
Follow-up to #916, which fixed the install pages. An audit of the rest of the current docs (
version-4.0.0,src/,content/llms-*.txt,docusaurus.config.js) found the same problems site-wide. This PR fixes them across 93 files.No editions, one Keploy
keploy.io/install.shand sign in withKEPLOY_API_KEY.ProductTier):/docs/server/installation/.eBPF on Linux, user space on macOS/Windows
Intro, concept and glossary pages, the docs home page's meta, the site schema, and the
llms.txt/llms-full.txttext (what AI assistants read) now say eBPF is how Keploy captures traffic on Linux. They no longer say Windows requires WSL2, and they give the native macOS/Windows scope:Quickstarts (31)
Each local-run section is classified by what the sample app talks to:
./mvnwcan't run undercmd.exe, and the Java classpath separator is;.Fixed along the way
keploy enterprise record(no such command) →keploy record --static-dedup.keploy.io/ent/install.share no longer offered for install.llms-full:keploy version→keploy --version;--skipCoverage→--skip-coverage.sudo docker compose up, and container names that don't exist in the sample.QuickStartList.js.codeRepository.Verification
npm run buildpasses.llms*.txtfinds no "Community Edition", "Keploy Community", "Keploy Enterprise", "Enterprise binary/edition/build",keploy enterpriseor "requires WSL".prettier@2.8.8 --check(the CI command) passes on every changed file.--skip-coverage,--freezeTime,--static-dedup,--custom-dedup-fields,--manual-login,keploy login,keploy --version).api-server/pkg/models/subscription.go.Needs an owner call (not changed)
src/schema/siteEntities.jsstill pointscodeRepositoryatkeploy/keploy. I removed the license claim, but the open-source positioning in keywords and topic tags is unchanged.EcosystemSupport.jskeeps "OSS" wording for the sponsorship programme for open-source projects.