CLI

terminal

danwo init
danwo test
danwo matrix tests/buy-credits.danwo.ts
danwo serve --open

danwo <command>; every command takes --help.

test

command

danwo test [patterns...]

run test files in worker processes

Member	Description
`patterns`	file paths or substrings narrowing the config's testMatch
`--lint`	run the integration lint over provider calls and webhooks (default: on)
`--no-lint`	skip the integration lint
`--reporter <names...>`	reporters besides the console's list: github (a check run, a pull request comment and the job summary in GitHub Actions), junit (.danwo/junit.xml)
`--no-pr-comment`	github reporter: no pull request comment, only the check run
`-w, --workers <n>`	worker processes
`--split-files`	hand workers single tests of the files whose tests share nothing (no beforeAll or afterAll)
`-r, --retries <n>`	retries per test
`-g, --grep <regexp>`	only tests whose full title matches
`--last-failed`	only the tests that failed in the latest run; merges into it
`--changed`	only the tests the changed files can affect (see danwo impact)
`--since <ref>`	what --changed compares with (default: HEAD, uncommitted changes)
`--shard <i/n>`	this machine's share of the test files, e.g. 2/4
`--max-failures <n>`	start no new test file once this many tests failed
`--max-duration <time>`	start no new test file once the run has taken this long, e.g. 20m (files left out fail the run)
`--ci`	CI defaults even without the CI variable: 1 retry, read-only cache, no .only
`--replay-only`	replay recorded agent steps and call no model: a step with no recording that replays fails
`--list`	print the picked tests as JSON (id, file) instead of running them
`-c, --config <path>`	config file (default: nearest danwo.config.ts)
`on-failure \| always \| off`	flight recorder: keep what runs did (default: config, else always)
`retain-on-failure \| on \| off`	video in kept records: retain-on-failure keeps failed runs' (default: config, else on)
`--keep-runs <n>`	how many of the newest runs keep their recordings (default: config, else 50)
`-u, --update-snapshots`	rewrite every toHaveScreenshot baseline

matrix

command

danwo matrix <file>

record a clean run of one test, then replay it once per planned fault

Member	Description
`file`	test file
`--lint`	run the integration lint over provider calls and webhooks (default: on)
`--no-lint`	skip the integration lint
`--reporter <names...>`	reporters besides the console's list: github (a check run, a pull request comment and the job summary in GitHub Actions), junit (.danwo/junit.xml)
`--no-pr-comment`	github reporter: no pull request comment, only the check run
`-g, --grep <regexp>`	selects the test when the file has several
`-f, --faults <regexp>`	only the planned faults whose labels match
`-w, --workers <n>`	worker processes (default: one per core, at most 8)
`--max-failures <n>`	start no new run once this many failed
`--shuffle <seed>`	run the faulted runs in a shuffled order
`--plan`	run the clean run, print the planned faults and how long they would take, and stop
`--budget <n>`	at most this many faulted runs (default: config, else 40)
`--all`	every fault on the test's body, background work included, with no budget
`--thorough`	every fault kind on each call, not only the most telling ones
`--every-occurrence`	faults on every repeat of a route or statement, not only its first
`--include-setup`	fault the test's setup too: its saved session signing in, fixtures and beforeEach
`-c, --config <path>`	config file (default: nearest danwo.config.ts)
`on-failure \| always \| off`	flight recorder: keep what runs did (default: config, else always)
`retain-on-failure \| on \| off`	video in kept records: retain-on-failure keeps failed runs' (default: config, else on)
`--keep-runs <n>`	how many of the newest runs keep their recordings (default: config, else 50)
`-u, --update-snapshots`	rewrite every toHaveScreenshot baseline

merge-reports

command

danwo merge-reports [files...]

join the results of a split run (danwo test --shard i/n) into one report

Member	Description
`files`	results files (default: .danwo/results/shard-*.json)
`--reporter <names...>`	reporters besides the console's list: github (a check run, a pull request comment and the job summary in GitHub Actions), junit (.danwo/junit.xml)
`--no-pr-comment`	github reporter: no pull request comment, only the check run
`-c, --config <path>`	config file (default: nearest danwo.config.ts)

impact

command

danwo impact <files...>

which tests a change to these files selects for danwo test --changed, and why

Member	Description
`files`	changed files
`-c, --config <path>`	config file (default: nearest danwo.config.ts)

mcp

command

danwo mcp

serve this project to coding agents over MCP (stdio)

Member	Description
`-c, --config <path>`	config file (default: nearest danwo.config.ts)
`default 4`	live sessions open at once

init

command

danwo init

set danwo up in this project: the engine, the model for agent steps, the providers to emulate and the coding agents to teach; writes danwo.config.ts, an example test and a test:danwo script, and installs what they import. Asks in a terminal; with --yes or no terminal it takes the flags and what it detects

Member	Description
`-y, --yes`	ask nothing: the flags, and the detected defaults for the rest
`web \| mobile \| both \| none`	what tests drive
`gateway \| anthropic \| openai \| google \| openrouter \| openai-compatible \| none`	where agent steps get their model
`--model <id>`	the model id, e.g. anthropic/claude-sonnet-5 on the AI Gateway
`--base-url <url>`	an OpenAI-compatible server's URL (--model-source openai-compatible)
`--key-env <name>`	the variable holding that server's key, when it takes one
`--providers <list>`	providers to emulate, comma-separated, or none (stripe, resend, s3, google, microsoft, slack, github, twilio, paystack, flutterwave, bachs, auth0, clerk, composio, nylas, nango, unipile, llm)
`--skills <list>`	coding agents to install danwo's skill for, comma-separated, or none (claude, cursor, codex, copilot, agents)
`--no-mcp`	do not register danwo's MCP server with those agents
`--dev-command <command>`	the command that starts the app, e.g. 'pnpm dev'
`--app-url <url>`	where the app answers once started
`--skip-install`	write the files but install no packages
`--force`	replace a danwo.config.ts that differs from the one init writes

doctor

command

danwo doctor

check what a first run needs: Node, danwo's Chromium, the config and its keys, the model's key, Playwright Test's version and the agents' skill

Member	Description
`--json`	the checks as JSON: { ok, checks: [{ name, status, detail }] }
`-c, --config <path>`	config file (default: nearest danwo.config.ts)

install-browser

command

danwo install-browser

install the Chromium build danwo drives, with danwo's own Playwright, once per machine; other Playwright installs' browsers are kept

Member	Description
`--with-deps`	also install the system libraries Chromium needs (Linux)
`--only-shell`	only the headless shell test runs use (smaller, for CI); not for danwo record
`--dry-run`	say what would be installed, and install nothing

record

command

danwo record [path]

record a test by using the app in a browser: locators for each action, what it caused behind the app, and a drafted invariant

Member	Description
`path`	the app page to start on
`-o, --out <file>`	the new test file, relative to the project (default: /recorded.danwo.ts)
`default recorded journey`	the test's title
`-c, --config <path>`	config file (default: nearest danwo.config.ts)

show

command

danwo show [run-id]

print the path of a recorded run's report (the newest by default)

Member	Description
`run-id`	a run under .danwo/runs
`-o, --open`	open the report in the browser
`html \| md, default html`	html: the report's path; md: a failure page for each failed run
`-c, --config <path>`	config file (default: nearest danwo.config.ts)

lint

command

danwo lint [run-id]

run the integration lint again over a recorded run (the newest by default); record passing runs too with --record always

Member	Description
`run-id`	a run under .danwo/runs
`-c, --config <path>`	config file (default: nearest danwo.config.ts)

serve

command

danwo serve

start the dashboard: runs, flight records, matrix grids and flaky tests from .danwo

Member	Description
`-p, --port <port>`	port on 127.0.0.1 (default 4310; 0 picks a free one)
`-o, --open`	open the dashboard in the browser
`-c, --config <path>`	config file (default: nearest danwo.config.ts)

cache

command

danwo cache <action>

list, count or clear the replay cache's recordings

Member	Description
`ls \| stats \| clear`
`-j, --journey <test-id>`	only the recordings of this test
`-c, --config <path>`	config file (default: nearest danwo.config.ts)

spec

command

danwo spec

print what this danwo provides, generated from its code: API, config, faults, codes

Member	Description
`--json`	compact JSON instead of text
`-t, --topic <topic>`	one section: api, fixtures, locators, expect, agent, secrets, config, environments, emulators, faults, cli, errors, cache, lint, exports

fidelity

command

danwo fidelity [providers...]

check the emulators against the real providers' sandboxes (with DANWO_FIDELITY_* credentials) or the committed goldens

Member	Description
`providers`	providers to check (default: all); `login google\|microsoft` signs a test account in instead
`--update-golden`	keep the sandbox's (or the API description's) structure as the new golden
`--from-openapi [file]`	make goldens from API descriptions: a file for one provider, or each provider's published one
`--offline`	compare against the goldens even when credentials are set
`--keep`	leave what live runs make in the sandboxes, listed as kept, instead of undoing it
`--sweep`	list leftovers tagged danwo-fidelity from earlier live runs instead of checking; --yes removes them
`--yes`	with --sweep, remove the tagged leftovers it lists
`--root <dir>`	the danwo repository holding fidelity/ (default: the current folder)
`--redirect-uri <uri>`	with login: the redirect URI the OAuth client registers (default: DANWO_FIDELITY__REDIRECT_URI, else http://127.0.0.1:8765/callback)
`--port <port>`	with login: the port of the default redirect URI

guide

command

danwo guide [topic]

list task guides, or print one

Member	Description
`topic`	a guide's topic; without one, the list

device

command

danwo device <action>

check this machine for mobile testing: Android (SDK, adb, emulator, rootable images, AVDs, JDK) and, on a Mac, iOS (Xcode, simctl, runtimes, CocoaPods)

Member	Description
`doctor`
`android \| ios`	check one platform only

build-android

command

danwo build-android [project]

build the danwo build of an Expo app as a release APK (trusting danwo's per-run CA); unchanged sources reuse the kept build

Member	Description
`project`	the Expo project's folder
`--force`	build again even when the sources are unchanged
`--verbose`	print the build tools' output

build-ios

command

danwo build-ios [project]

build the danwo build of an Expo app as a Release iOS Simulator app (plain HTTP allowed for the proxy); unchanged sources reuse the kept build

Member	Description
`project`	the Expo project's folder
`--force`	build again even when the sources are unchanged
`--verbose`	print the build tools' output

testCopy Anchor Link

matrixCopy Anchor Link

merge-reportsCopy Anchor Link

impactCopy Anchor Link

mcpCopy Anchor Link

initCopy Anchor Link

doctorCopy Anchor Link

install-browserCopy Anchor Link

recordCopy Anchor Link

showCopy Anchor Link

lintCopy Anchor Link

serveCopy Anchor Link

cacheCopy Anchor Link

specCopy Anchor Link

fidelityCopy Anchor Link

guideCopy Anchor Link

deviceCopy Anchor Link

build-androidCopy Anchor Link

build-iosCopy Anchor Link

On this page

test

matrix

merge-reports

impact

mcp

init

doctor

install-browser

record

show

lint

serve

cache

spec

fidelity

guide

device

build-android

build-ios