Test your apps in plain English.
On your infrastructure. Fully audited.
Orbit reads your screens the way a person does, so you describe what to check in plain language instead of wiring up brittle selectors. Built for regulated and financial teams: it runs entirely on-premise, every action is logged, and every result is reviewable.
157 capabilities · one on-premise platform · no selectors, no flaky scripts
Web testing
Full browser control driven by vision, from clicks and forms to dialogs, storage, and navigation.
Navigate to URLs
Go to any web address with a natural navigation command.
Click and tap
Click web elements by AI vision targeting or by coordinates, with no selectors.
Type text
Enter text into focused fields or AI-targeted positions without selectors.
Scroll
Scroll the page or a container up, down, left, or right.
Hover
Hover over elements to trigger tooltips, menus, and hover states.
Drag and drop
Drag an element from a source to a destination, including path-based moves.
Select from dropdowns
Pick a dropdown option using a plain-English description of the choice.
Double-click and right-click
Double-click elements or right-click to open context menus, all without selectors.
Press keys
Send keyboard keys such as Enter, Tab, and Escape for submission and navigation.
Upload files
Attach files to web forms through file inputs without selector dependencies.
Navigate history
Move back and forward in browser history during a test.
Refresh and reload
Reload the current page, clearing caches and re-running scripts.
Close window or tab
Close the active browser window or tab.
Click at coordinates
Click an explicit pixel position when you want to bypass vision targeting.
Native dialog handling
Auto-accept or dismiss alert, confirm, prompt, and beforeunload dialogs with per-action policies.
Permission granting
Grant geolocation, camera, notifications, and other browser permissions so platform features work under test.
Set cookies
Set browser cookies for authentication, session, and tracking state without touching the UI.
Set geolocation
Override device location to test location-aware features deterministically.
Storage management
Read, write, and clear LocalStorage and SessionStorage to test state persistence.
Mobile testing — native and mobile web
Drive real Android and iOS apps and mobile browsers with the same plain-English steps.
Launch and kill apps
Open an app by package name or bundle ID and force-stop it on demand.
Deep link navigation
Launch a specific app screen via deep link using a natural navigation command.
Tap, type, and swipe
Tap targets, enter text, and swipe in any direction, including swipes on specific elements.
Long press
Long-press at coordinates or on AI-targeted elements.
Double tap
Double-tap targets without selectors.
Home and back buttons
Press the device home button or navigate back during a test.
Dismiss keyboard
Hide the on-screen keyboard after text entry.
Native popup dismissal
Automatically detect and clear system popups such as permission, rating, consent, and ad dialogs.
Crash detection
Catch hard crashes and process death via Appium app-state checks, even on cloud farms where adb is unavailable.
API and database testing
Build, chain, and assert HTTP and SQL flows next to your UI steps — import from cURL, HAR, or OpenAPI.
HTTP request builder
Full HTTP step editor with method picker, URL input, and comprehensive request customization.
Headers and query parameters
Manage headers with smart suggestions and toggle query parameters per row with variable substitution.
Multi-mode request body
Send JSON, URL-encoded, form-data, binary, or raw text bodies with syntax awareness.
Authentication picker
Apply Bearer token or Basic Auth with credential input and automatic header generation.
HTTP settings
Configure SSL verification, timeout, and redirect following per request.
cURL import
Paste a complete cURL command to build an HTTP step, including method, headers, auth, body, and files.
HAR import
Import HTTP Archive captures from Chrome DevTools and bulk-create steps with filtering and selection.
Swagger / OpenAPI integration
Connect OpenAPI specs and pick operations to auto-scaffold HTTP steps with request templates.
Swagger spec and operation management
Add, refresh, and browse multiple specs, then search and filter operations before adding them to a flow.
Response variable extraction
Capture response values via JSON path, regex, or AI extraction into named variables.
Drag-drop response wiring
Drag a response field into a downstream header, body, query param, or URL to bind it automatically.
Response preview
Run a request once without saving a test, inspect the JSON, and drag fields out into variables.
Auto-wired API flows
Deterministically chain Swagger operations by matching path placeholders to response fields, no AI required.
HTTP assertions
Assert on status, headers, response time, JSON fields, full text, and variables with operators including semantic AI matching.
Database query builder
Author parameterized SQL against configured databases with variable substitution.
Database query preview
Run SQL against the real database with row-limit capping and inspect results before saving the step.
Database result extraction
Capture columns, row counts, and truncation flags from query results into variables.
Database assertions
Verify row counts, execution duration, truncation status, and individual column values.
Database safety policies
Block destructive operations like DROP, TRUNCATE, and ALTER, or restrict an integration to read-only.
Test data and datasets
Run any test or request across many rows, generate data with AI, and import from CSV.
Data-driven testing
Run a test once per dataset row, substituting column values into steps.
Inline dataset parametrization
Attach dataset rows to an HTTP request so it runs once per row with column substitution.
CSV dataset import
Paste CSV data into the dataset editor with automatic header parsing and row population.
AI dataset generator
Generate synthetic, type-checked dataset rows from a plain-English prompt and a column schema.
Dataset iteration overrides
At run time, swap a test's dataset, skip iteration, or run a single specific row.
Visual and UI checks
Catch layout, content, and rendering regressions with AI-assisted screenshot comparison.
Screenshots and visual regression
Capture screens and run AI-assisted visual regression checks across web and mobile.
Screenshot step builder
Add a visual check step in the browser with baseline selection and tolerance configuration.
Visual baseline snapshots
Create device-class-aware baselines automatically with an accept/reject management workflow.
Multi-device baseline isolation
Keep separate baselines per test, step, and device class so web, Android, and iOS rebaseline independently.
Baseline manager
Browse, organize, pin, refresh, and delete baselines grouped by test case and device platform.
Regression tolerance modes
Set per-step tolerance and choose strict, layout, or AI comparison sensitivity.
Ignore regions editor
Draw rectangular masks over dynamic areas like clocks and status bars to exclude them from comparison.
Visual diff viewer
Compare baseline, current, and difference side by side in a three-pane, color-coded view.
Per-check visual results
See per-step results for layout, spelling, image loading, and language checks with severity and AI reasoning.
AI visual narration
Get plain-English explanations of what changed in a failed comparison to speed up debugging.
Organize and scale
Group tests into suites with hooks, retries, scheduling, and environment control.
Test suites
Organize tests into reusable collections with shared environment, browser, strategy, and parallelism settings.
Suite lifecycle hooks
Run setup and teardown at the suite and per-test level with Before/After Suite and Before/After Each hooks.
Always-run teardown
Guarantee After-hooks run even when setup or a test fails, so cleanup always happens.
Sequential and parallel execution
Run tests one at a time in order or in parallel with a configurable slot count.
Stop on first failure
Halt a suite the moment a test fails to prevent cascading downstream failures.
Retries and re-run failed
Automatically retry failed tests with configurable attempts and delay.
Environment configuration
Define and switch between QA, staging, and production environments with environment-specific settings.
Global variables and secrets
Define organization-scoped variables, including secrets, with environment overrides across all tests.
Test prioritization
Assign Critical, High, Medium, or Low priority for risk-based ordering and triage.
Tagging and filtering
Tag tests with labels like smoke and regression, then filter the repository by tag, priority, platform, and metadata.
Session reset strategies
Control state between tests with fast reset, hard reset, or no reset to chain tests.
Suite scheduling
Schedule suites to run on recurring cron intervals within configurable time windows.
Reusable test contexts
Attach saved app-state and login-session snapshots to suites for reuse.
Execution and infrastructure
Run locally or across cloud device farms, with on-demand devices and full session artifacts.
Local device execution
Run on connected local Android and iOS devices or a local browser, with per-organization controls.
Cloud-only organization policy
Enforce cloud-only execution at the org level, transparently promoting local requests to the cloud.
Cloud device farm execution
Run native iOS and Android tests on BrowserStack, LambdaTest, and AWS Device Farm via Appium.
Multi-cloud dispatch
Route suites across multiple cloud providers and execute them in parallel.
Dynamic device pools
Acquire devices on demand at runtime for scalable parallel execution without pre-allocation.
Real device vs emulator detection
Distinguish real devices from emulators and simulators in cloud inventory and selection.
OS version constraints
Select devices by OS version using operators like equals, at-least, at-most, and latest.
Device capability mapping
Translate device selections into provider-specific Appium capabilities, including reset strategies and custom caps.
Cloud session artifacts
Automatically fetch Appium logs, network HAR files, device logs, and video recordings after each run.
App manager with cloud sync
Upload an APK or IPA once and sync it across LambdaTest, BrowserStack, and AWS with idempotent re-sync.
Web automation engine
Drive web tests locally and in the cloud via Playwright with browser, OS, and version selection.
Cloud provider registry
Plug in cloud device farms through an architecture that auto-discovers metadata, auth fields, and device lists.
Provider connection testing
Validate provider credentials and surface plan, parallel slots, and available devices during setup.
Cloud device listing
Query providers for devices with platform, OS version, manufacturer, tablet, and real/emulator metadata.
Session status reporting
Push pass/fail verdicts back to BrowserStack and LambdaTest dashboards in real time.
Queue-based distribution
Distribute runs across workers via a queue with parallel suite execution.
Rerun and stop controls
Rerun a test with identical settings in one click, or stop it immediately with a clean engine teardown.
Live test polling
See run status update in real time through periodic polling during active execution.
Run execution metadata
Track setup timing, first-step timing, and cloud session IDs for full run transparency.
Run artifact paths
Organize screenshots, logs, and videos under per-suite, per-run paths for retention.
File asset management
Upload and track general-purpose files for HTTP form-data and binary request bodies.
Reporting and audit
Drill into every run and step, with AI explanations, performance data, and an immutable audit trail.
Suite report
A redesigned suite-level report with a sticky filter bar, compact rows, and pass/fail/running/pending categorization.
Run viewer
A full run detail view with per-step drill-down, visual results, performance charts, and live polling.
Structured error classification
Categorize failures by taxonomy with human-readable labels and suggested fixes.
Per-step AI reasoning
Capture error type, analysis, suggested fix, and technical detail on each failed step.
Verification breakdown
Show per-assertion pass, fail, and reason for multi-assertion verify steps against a single screenshot.
Flow timeline view
Inspect HTTP steps with method, URL, and status alongside request, response, and extracted-variable detail.
Performance report panel
Chart CPU, memory, JS heap, FPS, network, and Core Web Vitals with budget-violation alerts.
Console events
Capture browser console logs and page errors during each step with timestamps.
Network events
Capture per-step HTTP requests with method, URL, type, status, size, and duration, with filtering.
Long tasks and animation frames
Detect jank with a Long Animation Frame and Long Task timeline showing blocking scripts.
Dialog handling capture
Record native dialogs with message, action taken, and timing.
Performance monitoring
Assert on Core Web Vitals, long animation frames, and network latency during execution.
Iteration navigation and data
Step through data-driven iterations with an N-of-M counter and per-iteration dataset values.
Per-step duration metrics
Show wall-clock duration per step, excluding queue wait, in human-readable form.
Extracted variables display
Render step-level variables extracted from HTTP responses and browser runs for chaining visibility.
Device and provider info
Display execution environment metadata including device name, provider, and platform.
Run cost tracking
Break down total cost across cloud execution, LLM usage, and manual QA per minute.
Analytics history
Query execution history filtered by version, environment, trigger, platform, and status, with error-type counts.
Artifact storage and download
Browse a hierarchical suite-and-run tree with per-run URLs and streaming download.
Screenshot baseline API
List, accept, update, and delete visual baselines via REST with project-wide and per-test views.
Platform audit log
Keep an immutable trail of admin actions with IP, user-agent, and payload for compliance review.
Platform, security, and deployment
On-premise by design, with tenant isolation, encrypted secrets, and white-label branding.
On-premise deployment
Deploy the whole platform via multi-container Docker Compose — frontend, backend, queue worker, database, and Redis.
Auto-SSL reverse proxy
Route traffic and auto-provision HTTPS certificates via Caddy for on-premise and cloud deployments.
Local artifact storage
Keep artifacts on the local filesystem for full data residency, with no external dependency.
Cloud storage backends
Optionally use pluggable S3 or GCS artifact storage with per-organization configuration.
Organization tenant isolation
Scope every endpoint by organization, with org-id claims verified per request and suspension checks.
Granular permissions
Control access with admin and member roles plus per-resource permissions stored per user.
Encrypted credential storage
Encrypt sensitive tokens — database DSNs, OAuth secrets, auth headers — at the database column level.
DSN masking
Strip passwords from database connection strings in the UI while keeping user and host visible.
API token management
Create, list, and revoke scoped API tokens for CI triggering and test generation.
Public test generation API
Generate tests programmatically over REST with token auth, scope checks, and PRD or text input.
CI/CD pipeline trigger
Kick off suite execution from a pipeline using API-key header authentication.
Database integration vault
Manage multiple encrypted SQL connections — PostgreSQL, MySQL, MSSQL — with a test-connection UI.
JIRA issue tracking
Auto-file bugs to JIRA with failure context from suite runs.
GitHub integration
Sync PR statuses and commit-level test results to GitHub via OAuth and webhooks.
Integrations hub
Manage 15+ integrations across source control, test infrastructure, storage, and notifications in one dashboard.
White-label branding
Theme product name, logo, colors, and domain per deployment, with build-time injection requiring no code changes.
See these features run on your real app
We'll author and run a test on your own application — on-premise, audit-ready, and live while you watch.