What you can build with Flowstride
Flowstride (FS) brings browser actions, APIs, reusable data, assertions and evidence into readable .flow scenarios. Use the same suite interactively and in CI. Use the separate Manual Workspace for cases that need a person's observation and judgment.
FS runs locally without an account. A free Cloud account connects workspace services; access and usage depend on your workspace plan. Cloud is an account-connected engine and service layer, not a requirement to host the application or surrender control of your CI runner.
Choose the outcome you need
| What you want to verify | Flowstride workflow | Learn it |
|---|---|---|
| Every price in a dynamic catalog | Calculate discounts in JS and match all 50 browser cards by SKU in TS | Complex tests with helpers |
| A sign-in that depends on another test | Clean up failed attempts and hand off the account after its role is saved | Fixtures and handoffs |
| A complete customer journey | Browser actions, HTTP/GraphQL preparation, extraction and shared data in one scenario | API + UI |
| Signup and email verification | Generate an address, retrieve a delivered FlowMail OTP, enter it and assert successful login | Email OTP |
| A control with repeated text | Element qualifiers, spatial anchors, indexes and explicit selectors | Targeting |
| Forms and dependent choices | Text input, native/custom selects, date setting, checkboxes, uploads and dialogs | Forms and dates |
| A record in a large table | Search, identify a row, act inside a cell, extract a value and reuse it elsewhere | Tables and cells |
| An embedded page or popup | Readable iframe actions, frame-scoped assertions and tab switching | Tabs and iframes |
| A recording and its transcript | Feed a WAV microphone fixture, drive recording controls and check the application's transcript | Audio-to-text |
| Two independently logged-in users | Named browser profiles with separate cookies, storage, tabs and permissions | Independent users |
| Repeatable login across runs | Save and restore cookies, localStorage and sessionStorage | Saved sessions |
| Reliable, precise checks | Retrying conditions with individual target/condition results and expected/actual diagnostics | Assertions and failures |
| API contracts | Structured HTTP requests, GraphQL, named clients, JSON Schema, polling and typed extraction | API and network |
| Slow, failed or controlled services | Browser route fulfillment, patching, request matching and delays, plus armed request/response waits | Network control |
| Reusable setup and test data | Load files, repeat a scenario per row, seed generators, reuse fragments and register cleanup | Files, data and run control |
| Inspect a run or choose which scenarios run | Logs, deliberate waits, dashboard pause, pacing, tags and recorded skips | Execution controls |
| Application behavior beyond the page text | Console messages, WebSocket and SSE events, visual baselines and diff evidence | Application evidence |
| A generated report or export | Arm a download, save it and check filename, size, content, JSON or hash | Downloads |
| Time-dependent behavior | Page clocks, controlled timers, function spies and return/throw stubs | Time and functions |
| Location, connectivity or browser settings | Runtime permissions, geolocation and offline recovery; profile-specific settings | Browser environments |
| A reusable UI component | Mount, interact, update and unmount vanilla, React or Vue components | Component testing |
| Accessibility regressions | axe scans, ARIA trees and reviewed local baselines | Accessibility |
| Rendering and game rules | Canvas input, state fixtures, screenshots, pixel comparisons and Three.js selection | Browser games |
| A failure in a previous attempt | Per-file screen history, attempts, network logs, Replay, traces and HAR | Dashboard · Evidence |
| Repeatable pipeline execution | Native CI, cancellation, retries, GitHub workflow generation, browser matrices, sharding and merging | Native CI |
| A human-executed test | Canvas design, grouped steps, expected outcomes, statuses, defects, media and retained runs | Manual testing |
Start with a real example
Browse every structured command or go directly to all structured assertions for one example per browser condition, stored-value operator and specialized check.
Run the examples included in your npm installation. They provide a local application and known passing/failing checks for API + UI, iframes, tables, variables and saved sessions. There is also a microphone fixture and a manual design. Service templates explain what your application must supply for real OTP delivery or speech recognition.
The guides distinguish a runnable example from an application contract you must adapt. A passing recording fixture demonstrates microphone capture and assertions; a real speech service must still recognize the words. A successful click establishes the action; your assertion establishes the business outcome.
Extend a focused action
Import a named JS or TS helper for loops, branching, calculations or work with the active Playwright page. A TypeScript plugin can use the current browser page, HTTP helpers, state and attachments. Keep suite orchestration, CI status, browser cleanup and evidence ownership in Flowstride. Capabilities and boundaries records the supported limits.
