Skip to content

Manual capabilities and current boundaries ​

Use this guide to plan manual testing around the capabilities available in Flowstride and Flowstride Pro modes of the local workspace. Cloud includes free accounts; connected services remain subject to workspace limits.

AreaCurrent behaviorPractical boundary
CanvasSteps, groups, text, connections, zoom, selection, undo/redo and internal clipboardIn-memory editing history is not backup/version control
Case detailsDescriptions, steps, expected results, references, priority and flow metadataNo dedicated typed field for every test-management concept
ExecutionHuman judgment, connected-path progression, six step statesNot conditional branching or automated application control
StorageLocal templates, submitted run JSON and media directoriesSave confirmation matters; local files are editable
FS saveServer persistence; an additional toolbar export path depends on the hosting bridgeDownload wrapper is not the same format as a submitted run
Cloud mode historySaved report review with historical editing/toolbar locksNot a resumable or cryptographically immutable run
CloudAuthenticated run synchronization where eligibleNot live collaborative canvas editing
AssigneesDescriptive values attached to defectsNo notification, directory validation or issue creation
EvidenceLocal reference/defect media; remote defect upload when availableNew Cloud attachments require workspace access; historical links need separate review; local references do not travel automatically
ReportsSaved step-level report and eligible Cloud reviewAutomated TestCase PDF is not a promised manual PDF export
AutomationA manual case can guide .flow authoringNo native one-click converter

Edition-specific launch behavior ​

The FS open command launches the local workspace without authentication. Cloud mode checks the account session before opening the local workspace. Open the workspace from your project:

sh
npx flowstride open

Both have local project APIs. Credentials available to a manual server can enable synchronization; this is separate from native CI's cloudSync switch. Do not assume that setting the automated CI option disables every manual-server Cloud request.

Plan around current limits ​

Execution follows the first outgoing connection, starting with the lowest internal node ID. Use separate linear flows rather than conditional branches or loops. Display numbers do not reorder internal IDs.

The defect chooser advertises a larger video limit than the local upload server accepts. The effective request cap is 40 MiB. Aggregate pending counts include Blocked and Skipped work. Client-provided author/tester labels do not prove identity.

Keep these limits in mind when planning coverage and reviewing saved outcomes.

Local access and safe sharing ​

The current manual server binds to loopback by default and validates local request origins/hosts. Imported content and save payloads are validated and rich HTML is sanitized. These controls do not turn the server into a public team-hosting endpoint. Keep it local and share reviewed run files or eligible Cloud records through their documented channels.

For a complete exercise with a real local application, a passing run and a deliberate mismatch, use the profile-review walkthrough.

Flowstride is MIT licensed. Flowstride Cloud offers free and paid workspace plans.