Features
Every step of the workflow, covered
From persistent browser profiles to a loopback-only API, here is what each part of ProtoX Login does — and what it guarantees about durability, isolation, and control.
Set up
Profiles and browsers
Each profile keeps its own cookies, sessions, and extensions in place between runs, so logged-in work continues where it left off. The browsers themselves are managed as well: installed from built-in sources, validated, and versioned per profile.
- Create, clone, and edit profiles — each with a stable ID and its own user-data directory, isolated from every other profile.
- Install browsers from built-in sources; versions sit side by side and each profile pins one validated version.
- Per-profile settings for proxy, locale, timezone, viewport, color scheme, user-agent, and startup pages.
- Open any profile for manual use in a real browser window, with the same isolation as automated runs.
Capture
Recorder and selectors
Record your own clicks and typing into editable workflow steps, then fine-tune them in the editor. Selector previews show exactly what each step will match before you commit to it.
- Record navigation, clicks, form input, keyboard, scrolling, and page waits into editable steps.
- Pick elements directly on the page, including inside frames, and inspect their selectors.
- Selector preview shows matches, frame path, ambiguity, and failure reasons before you save a step.
- Selectors are built on Playwright locators, so they hold up as pages shift.
Run
Execution and validation
A run follows one deterministic path through the graph. Validation before execution, checkpoints along the way, and error handles on every node mean failures are precise and visible — never mysterious.
- Conditions, loops, nested workflows, delays, retries, timeouts, and per-node error handling.
- Execution state is checkpointed at node boundaries, so interruptions stay visible and recoverable — never silently replayed.
- Variables, secrets, and previous-node output flow into later steps as runtime data.
- Scripts run sandboxed by default; trusted mode is an explicit choice you make per workflow.
Automate
Scheduling and queue
Put routine work on cron-like schedules and let the durable queue handle ordering, retries, and missed runs. The scheduler keeps going even when the window is closed to the system tray.
- Manual, one-time, interval, cron, app-startup, chained, and API triggers for every routine.
- Schedules carry a timezone and a missed-run policy: skip, run once, or bounded catch-up.
- The queue is partitioned by profile and ordered by priority, with one active run per profile.
- Jobs are persisted before they are acknowledged, so a restart never loses queued work.
Debug
History, artifacts, and diagnostics
Every run is recorded: structured logs, node results, checkpoints, and artifacts on disk. When something goes wrong, the health view and diagnostic bundles point at the actual cause.
- Inspect queued, running, completed, failed, cancelled, and interrupted executions in one place.
- The live view shows the current node, profile, browser session, duration, and recent events.
- Artifacts — screenshots, downloads, exports, traces — are kept on disk with configurable retention.
- Errors carry a stable code, technical detail, and a suggested recovery action.
Integrate
Local API
A REST API on localhost connects ProtoX Login to the rest of your tooling. It is local-only by default and secured with scoped tokens, so automation stays under your control.
- Trigger runs, check status and results, and cancel — from your own scripts and tools on the same machine.
- Bearer tokens with scopes, rotation, and revocation.
- Idempotency and correlation IDs keep retried requests from creating duplicate runs.
- Binds to 127.0.0.1 only, and never exposes raw browser debugging or the filesystem.
Ready to automate?
Download ProtoX Login for Windows and run your first workflow in minutes.