Pricing
Pay per usage
Drive browser actions and replay existing scripts in various formats or convert agentic actions into code (future)
Pricing
Pay per usage
Rating
0.0
(0)
Developer
Vasek Codey Vlcek
Actor stats
1
Bookmarked
2
Total users
1
Monthly active users
a day ago
Last modified
Categories
Share
username ➜ login.usernamepassword ➜ login.passwordproxyConfig ➜ proxyCleanup legacy bits. Refactors, minor fixes.
Rewrite of notifications. Rewrite of loggers & event handlers to dynamic solution to remove dependency on static definitions. Target url no longer required in setup.
Refactoring & cleanup. Minor fixes.
Robot.Setupoptions.output.filter - add option to filter out nullish values from the outputRefactoring & cleanup. Update example implementation.
ESM hotfix + upgrade of major dependencies
Robot.SetupInterface server is no longer bound to page events out of the box and will not update the interface automatically, unless the new options are enabled:
options.interface.events.serveOnEvents - enable automated updates of the remote interface based on specific page events
options.interface.events.eventHooks - events to trigger automated updates of the remote interface (e.g. waitForSelector)
Fix: Provide server screenshot before closing browser.
Provide support for custom (sticky) proxy sessions using Apify sessionPool
Update of major dependencies. API remains unchanged.
Custom views for visual interaction through interace server. Improved support for frames both in public tools and internally. Internal improvements and cleanup of automated logging and debugging tools and hooks. Automated response error (>400) logging moved to debug mode only to minimize risk of unintended leaks.
Robot.Setupoptions.interface.client - support for custom client views and interfaces on the interface serverRobot.toolsintegrateInstance - integration of custom page or frame instances at runtime to enable on them the built-in automated facilities of the framework (support for frames)Internals are currently backward compatible to enable a smooth transition from older versions, the upgrade will however require at least an update of a project's Robot.Setup and any tools updated in the "breaking" section. Ideally, input schema and API of the implementations should also be updated. Legacy functionality and backward compatibility will be completely removed in upcoming versions in the future.
process.env.slackToken ➜ process.env.SLACK_TOKEN
retry - retry will now trigger only on custom errors with the retry flag set to truesilent - removed in favor of replacement options below (notify, options.debug.muted)RobotRenamed variables for more clarity, merged actor and robot input and output. Input for the robot is no longer clearly separated from actor input and instead merged and flattened with the rest of input due to input schema constraints and considerations. Workaround for this if it should ever become an issue is pre-processing of the input before passing it to the robot for processing and accessing nested properties during automation as pre-defined. This change also rules out the potential for processing multiple inputs within a single run without the need for a higher order management actor but since the use case is unlikely and would introduce significant overhead with less clarity in terms of managing and matching inputs and outputs, drawbacks seem to far outweight the potential benefits and thus this kind of usage will not be supported in the future.
INPUT + input ➜ inputOUTPUT + output ➜ outputRobot.SetupOutputTemplate ➜ OutputSchemagetApifyProxySession ➜ getProxySessionIdOPTIONS ➜ optionsOPTIONS.blockRequests ➜ options.trafficFilterOPTIONS.blockRequests.patterns ➜ options.trafficFilter.patterns.urlOPTIONS.blockRequests.analytics ➜ options.trafficFilter.patterns.hostSERVER ➜ options.serverSERVER.liveView ➜ options.server.interfaceSLACK.channel ➜ options.notify.channels.slack.channelRobot.Scope/Targetconstructor(setup, target, robot) ➜ constructor(context, [robot]) - simplify optional scope/target constructors
stepName = () => {} ➜ taskName = context => ({ stepName: context => {}, ... }) - steps are now optionally (recommended and will become default) wrapped by their respective tasks in Robot.Scope/Target implementations to follow the structure in Robot.Setup more closely and provide other additional benefits through this closure, including type hints in step signatures. Updated context is passed to both tasks and steps at runtime.
context
relay ➜ stateINPUT + input ➜ inputOUTPUT + output ➜ outputRobot.constsPUPPETEER.events ➜ EVENTSRobot.tools.loginselectors.loggedIn ➜ selectors.verifypredicate or selectors.verify is present for login status verificationInputbrowser - select from available browsers to perform the automationhuman - enables optional human behavior simulation tools in contextnotify - enable error notifications to external channel(s - future)prompt - enable user prompt and require manual intervention before proceeding with the automation (currently must be implemented in scope/target)options
debug
pwApi - very verbose logs of the automation library (Playwright only)pwAll - extremely verbose logs of the automation library (Playwright only)muted - mute error notifications to external channelstraffic
enable - enable realtime traffic logs (very verbose)fullUrls - log full URLs including complete parametershostOnly - only log traffic with the target host domainhideFilter - mute traffic blocked by traffic filters in logsnotify
details - include complete error in error notificationvisuals - include visual indicator of important input optionsverbose - include messages from the automation libraryslack - enable channel for error notifications (Slack)proxy
proximity
enable - enable proxy selection based on geolocation approximation based on input IPserver
interface
enable - enable visual interface for live inspection of and interaction with the automationwebsocket
enable - enable built-in websocket server for external real-time network communication (future)Robot.CaptchaSolverAutomated captcha solver utilizing the paid services provided by AntiCaptcha
getSolution - manual handling of the captcha solution received from servicesolveCaptcha - automated handling of the captcha including injection attemptinjectSolution - attempt to inject captcha solution at target automaticallyRobot.ErrorNative support for custom errors with special flags reserved for use by the robot and support for JSON.stringify
Non-custom native errors are wrapped automatically, though only shortly before error alert and exit of the actor.
Exhaust options: throw new Robot.Error({name, type, data, error, retry, message, rotateSession, retireSession})
Rethrow wrapped error:
throw new Robot.Error({error})cause of re-thrown error throw new Robot.errors.Example({error, data: {test: 'anything'}})Reserved properties:
data: Object - container for arbitrary data propertiestype: String - constructor name of the custom errorerror: Error - optional cause of the custom errorretry: Boolean - retry current run till retry limitrotateSession - rotate proxy session on retryretireSession - rotate proxy session on retryRobot.HumanEnable humanized toolset for manual or automatic simulation of human behavior. Enable automatic pointer tracking and visual movement indicator in debug mode.
human
click - humanized clicking with random movements and action delayspoint - humanized pointing with random movements and action delayspress - humanized keypress with random movements and action delayssleep - humanized waiting with random movements and action delaystype - humanized typing with random movements and action delaysAutomatic spontaneous motion is currently disabled by default. Manual usage:
human.startMotion - start automated continuous simulation of human behaviorhuman.stopMotion - stop automated continuous simulation of human behaviorRobot.SetupDocumented in-line in the base class Robot.Setup
step.abort(context) - support centralized abort trigger across all targetsoptions.browserPool - browser pool options with extra options and mappingsoptions.sessionPool - session pool options with extra options for the robotoptions.sessionPool.disable - disable session pooloptions.browser - browser utilized for automationoptions.library - lower level automation library optionoptions.launchContext - options for standalone browser (no stealth support)options.notify.details - include extra error details in external notificationsoptions.notify.filters - specify error alert filters based on error name/typeoptions.notify.visuals - include visual indicator of important input optionsoptions.server.interface.prompt.handlers - handlers for interface promptoptions.trafficFilter.resources - filter out requests by resource typeRobot.ScopeIntroduce generic scope abstraction for larger, target independent automations
Robot.TargetExtends Robot.Scope
Robot.Scope/TargetConstructor parameters:
context - context passed to every instance of Robot.Scope/Targetrobot - back-door to the running robot instance for more advanced runtime manipulation (use with caution!)Bindings between the contents of context and instances of this class are automatically initialized through the base class. Therefore constructor is optional if there is no need to manage properties or local state of the scope. For sharing and managing global state across all scopes at runtime, use the state object hosted in context.
context
task - task currently being processedstep - step currently being processedinput - input as defined by input schema
options - input options for various robot features transformed into an object at runtime from the flat input schema using dot notationoutput - current output of the automationpage - instance of browser page (Puppeteer/Playwright)
gotoDom - utility method for navigation with waitUntil option enabled and set to EVENTS.domcontentloadedevents - future tools for emitting and listening to custom eventshuman - optional human behavior simulation tool to replace native methods where neededpools - instances for utility management pools of the Apify SDK
BrowserPool - exposed browser pool instance if applicableSessionPool - exposed session pool instance if applicabletools - convenience automation tools preloaded with page
debug - capture named debug buffers to store on demandmatchPattern - pattern matching utility for a single patterniteratePatterns - pattern matching utility for multiple patternsstate - utility relay object for passing state and data between steps and scopesserver - general purpose server for real-time network communication and visual interaction with the actor
interface - visual interface for live inspection of and interaction with automation runtime
start - initialize and launch the interface server (automatic with input option enabled)serve - capture a snapshot of current browser state and update the visual interfaceprompt - prompt an automation action with visual context through the interfacehypertext - future universal hypertext server with dynamic custom handlerswebsocket - future universal websocket server with dynamic custom handlersProperties available directly on the scope instance, outside of context:
this.debug - capture named debug buffers to store on demandthis.task - currently processed task as defined in Robot.Setup
this.task.output - output of current taskthis.step - currently processed step as defined in Robot.Setup
this.step.output - output of current stepthis.step.attachOutput(output: object) - attach output to current step (object merge)this.step.output.attach(output: object) - same as abovethis.step - current global output is available outside of the context, directly on the scope instancethis.will(text: string) - trigger a visually highlighted inline step without a separate step functionstep.will(text: string) - same as above for generic or loose steps outside of Robot.Scope/Target
Convenience shortcuts already preloaded with page
this.matchPattern(pattern) - curried version on Robot.Scope/Targettools.matchPattern(pattern) - curried version in Robot.context.toolsthis.iteratePatterns(patterns, [patternOrder]) - curried version on Robot.Scope/Targettools.iteratePatterns(patterns, [patternOrder]) - curried version in Robot.context.toolsRobot.errorsLimited dictionary of the most common errors with support for error instance matching. Default errors that can be further modified or extended for the needs of a specific automation project.
Robot.errors.Network - throw, print and report a generic network errorRobot.errors.access.Blocked - throw, print and report a generic access errorRobot.errors.session.Rotate - rotate proxy session before retrying a failed actionRobot.errors.Status({error, retry: true, retireSession: true, statusCode: 403}) - rethrow previous error as cause of the custom error, retire proxy session before retrying failed action and print a message with failed status codeRobot.toolsPattern matching has been updated to only match elements actually visible on the page in order to avoid false match positives with hidden elements not intended for current state of the user-facing website interface.
debug(page, name) - store debug buffers on demandmatchPattern(page, pattern) - match a single exact patterniteratePatterns(page, patterns, [patternOrder]) - test multiple or all patterns until a match is foundINPUTdebug - enable debug mode but no longer silence error alertssilent - enable silent mode (mute external notifications)Robot.toolsverifyResult(selector, contents) ➜ verifyResult({selector, contents})
searchPolicyNumber ➜ searchResult
INPUTdebug + retry - debug mode will save debug buffers to store on each retry