Settings

Introduction

What is the Settings screen?

The Settings screen (route /settings) is Testver’s central control panel. From a single page you configure how the application looks, which testing surfaces are visible, exactly how your tests are launched on the command line, which AI providers and models power Testver’s intelligent features, and which live browser processes are currently running on the server. Everything here is scoped to your project and persists between sessions.

Note

In one sentence: Settings is where you tell Testver about your project’s run commands, connect your AI providers, choose your theme and workspace mode, and watch over active browser sessions.

Who this guide is for

• QA engineers and testers who need Testver to launch their existing test suite correctly (single test, whole file, full run).
• Developers wiring up AI features who need to add an API key, register models, and pick a default provider.
• Team leads / workspace admins who set the App Mode (Web vs Mobile) and manage browser sessions across all users.
• Anyone who wants to switch between the light and dark theme or check basic project information.

Key terms

Term	Meaning
Provider	An AI service that supplies language models — e.g. OpenAI, Anthropic, Google AI, xAI, or a local server like Ollama / LM Studio. Configured on the LLM tab.
Model	A specific named AI model offered by a provider, such as gpt-4o, claude-sonnet, or gemini-2.0-flash. Each provider holds a library of models, one of which is marked active.
API key	A secret credential that authorizes Testver to call a cloud AI provider. Entered per provider on the LLM tab; cloud providers require one, local providers do not.
App mode	A workspace-wide setting (Web or Mobile) that controls which Testver surfaces — Runner, Recorder, Schedules, AI Assistant — are shown. Set on the General tab.
Run command	The exact shell command Testver uses to execute your tests. There are three: Single Test, Suite / File, and Full Run. Defined on the General tab under Run Commands.
Env var (environment variable)	A KEY=value pair injected into the environment when your tests run. Used for things like credentials or flags; managed under Run Commands → Advanced Settings.

How settings take effect

• Most settings are saved per project on the Testver server and survive restarts.
• Run Commands are only stored as an override when you change them away from the auto-detected default — otherwise Testver keeps using its detected command, so future framework changes are still picked up automatically.
• LLM provider settings (API keys, model libraries, base URLs, timeouts) are saved per provider when you click Save. The default provider can also come from an environment variable instead of being typed in.
• App Mode is workspace-wide: changing it reloads Runner, Recorder, Schedules, and AI Assistant for every user in the workspace.
• Theme (light/dark) is applied instantly and remembered locally in your browser.
• Browser Sessions are live server state, not saved configuration — the list refreshes automatically every few seconds.

Note

Saving is per-card. Each section (Run Commands, each LLM provider, etc.) has its own Save button — clicking Save on one card does not affect unsaved edits on another.

Getting Started

Opening Settings

1. Open Testver and make sure a project is loaded.
2. Click the Settings entry (gear icon) in the navigation, or browse directly to /settings.
3. The page opens on the General tab by default. The current tab is reflected in the URL via the tab query parameter (for example /settings?tab=llm), so you can bookmark or share a direct link to a specific tab.

The Settings tabs

Three tabs run across the top of the Settings screen. Click any tab to switch; the active tab is underlined in the brand color.

Tab	Icon	What it covers
General	Palette	Project info & appearance — App Mode, Project Information, Run Commands, and the theme toggle.
LLM	Brain	AI provider configuration — API keys, model libraries, default provider, base URLs, and timeouts.
Browser Sessions	Server	Active browser processes — view and close server-side, local, and test-run browser sessions.

At a glance

• Start on General to confirm your project details and make sure your test run commands are correct.
• Move to LLM to plug in at least one AI provider if you want Testver’s AI-powered features (such as Suggest with AI for run commands).
• Use Browser Sessions when you need to see or shut down browsers Testver has launched.
• Use the theme toggle on General any time to switch between light and dark mode.

General Tab: App Mode

The App Mode card sits at the top of the General tab. App Mode controls which Testver surfaces are visible across the entire workspace — the Runner, Recorder, Schedules, and AI Assistant pickers all adapt to this setting. The card shows when the mode was last changed.

Option	What it does
Web tile	Sets the workspace to web testing only — shows just the web testing UI. Accented in sky blue.
Mobile tile	Sets the workspace to mobile testing only — shows just the mobile testing UI. Carries a Beta badge and an indigo accent.
Confirmation dialog	Clicking a different tile does not switch immediately. A Change App Mode? dialog appears explaining the impact before you confirm with Switch to <mode> (or Cancel).
Last changed date	Displayed at the top-right of the card when the mode has been set before, showing the date of the most recent change.

Caution

Switching App Mode affects the entire workspace. Runner, Recorder, Schedules, and AI Assistant reload and show only the new mode’s controls, and active recording or test-run sessions tied to the old mode may become unreachable from the UI. Saved schedules and history remain on disk and reappear if you switch back.

Project Information

The Project Information card is read-only. It summarizes what Testver has detected about your project so you can confirm it loaded the right codebase.

Field	What it shows
Framework	The detected test framework, or Auto-detected if not explicitly set.
Language	The project’s primary language, or Multiple when several are present.
Path	The absolute root path of the project on disk (shown in monospace).
Modules	The total number of modules Testver found.
Total Files	The total number of files scanned in the project.

Run Commands

The Run Commands card defines exactly how Testver launches your tests. Testver auto-detects sensible commands for your framework (shown as a badge next to the heading) and pre-fills them. You can edit any of the three commands, test them, reset them to the detected default, or have AI suggest them.

There are three command fields, each in its own panel with Reset and Test buttons and an Auto-detected / Custom badge:

Command field	What it does
Single Test	Runs a single test/scenario by name.
Suite / File	Runs all tests in one file or suite.
Full Run	Runs the entire test suite.

Each field’s controls:

Control	What it does
Custom / Auto-detected badge	Shows whether the current value differs from the detected default (Custom) or matches it (Auto-detected).
Reset	Restores the field to Testver’s auto-detected default command.
Test	Dry-runs the command (placeholder tokens are substituted with harmless dummy values) and reports Command is valid or Command failed, with a snippet of the output.
Command input	The editable shell command, shown in monospace. The placeholder text is the auto-detected default.

Placeholder tokens — commands may include tokens that Testver replaces at runtime with the real test name, file path, and so on. Each token chip on the card is click-to-copy. The available tokens:

Token	Replaced with
{test}	Test name
{suite}	Suite name
{line}	Start line number
{args}	Extra arguments
{file}	Relative file path
{absFile}	Absolute file path
{posixFile}	Forward-slash path (useful for WebdriverIO —spec on Windows)
{fileName}	File name
{fileNoExt}	File name without extension
{project} / {cwd} / {root}	Project root directory

Suggest with AI — the button beside the Run Commands heading lets AI inspect your project and propose all three commands.

Element	What it does
Suggest with AI button	Sends your project to the AI, which analyzes it and returns proposed Single Test, Suite/File, and Full Run commands. Shows Analyzing project… while it works.
Suggestion cards	Each proposed command shows a confidence percentage, a validation badge (checking / resolves / unverified), the command in monospace, and a short rationale.
package.json script badge	Appears when the suggestion reused an existing npm script from your project.
Apply (per card)	Fills that one field with the suggested command (review and Save to persist).
Apply all	Fills all three fields with the suggestions at once.
Discuss in AI Assistant	Hands off to the full AI Assistant with a context-rich seed prompt for complex cases (does not auto-submit).
Dismiss (X)	Closes the suggestion panel without applying anything.

Note

Suggested commands are not saved until you click Save Run Commands. Applying a suggestion only fills the field — review it first.

Advanced Settings

Clicking Advanced Settings expands extra options below the three command fields.

Option	What it does
Environment Variables	Add any number of KEY / value pairs that are injected into the environment when your tests run. Use Add Variable to add a row and the X button to remove one. Empty rows are ignored on save.
Before Command	A shell command run before your tests (for example npm run build).
After Command	A shell command run after your tests (for example npm run cleanup).

Click Save Run Commands at the bottom of the card to persist everything. Testver only stores a command as a custom override when it differs from the auto-detected default; the working directory, before/after commands, and non-empty env vars are saved as entered.

Appearance & Theme

Testver supports a light and a dark theme. The theme toggle flips between them, and your choice is remembered locally in the browser and applied instantly across the app — no save button required. The current theme is read on load so Settings always reflects what you last picked.

LLM Providers & Models

The LLM tab lists every AI provider Testver supports as an expandable card. Each card shows the provider name, a short description of its models, a status indicator, and a star to mark it as the default provider. Click a card header to expand its configuration form.

Provider	Models / role	Needs API key?
OpenAI	GPT-4o, GPT-4, GPT-3.5 Turbo	Yes
Anthropic	Claude Sonnet, Claude Opus, Claude Haiku	Yes
Google AI	Gemini 2.0 Flash, Gemini Pro	Yes
xAI	Grok-2, Grok-3	Yes
Ollama	Local models — Llama, Mistral, Phi, etc.	No (local)
LM Studio	Local model server	No (local)

Provider status — the colored dot and label on each collapsed card tell you how the provider is configured:

Status	Meaning
Not Configured (gray)	No key or settings saved for this provider yet.
Configured (green)	An API key has been saved in Testver’s config.
Env Var (green)	The API key is supplied via an environment variable rather than typed into Testver.
Local (amber)	A local provider (Ollama / LM Studio) that runs models on your own machine.

Default provider — the star icon on each card header sets that provider as the workspace default. The current default shows a filled amber star and a DEFAULT badge. Clicking the star on another provider switches the default immediately.

Configuring a provider

Expanding a provider card reveals its full form. The exact fields depend on whether the provider is cloud (needs a key) or local.

Field	What it does
API Key	The secret key authorizing calls to the provider (cloud providers only). Hidden by default; use the eye toggle to reveal it. A saved key shows as dots with a (saved — click to change) hint. If the key comes from an environment variable, a note names the variable instead.
Models	The provider’s model library, shown as chips. Click the star on a chip to make it the active model; click the X to remove a model (when more than one exists).
Add model input	Type a model name and press Enter or click + to add it to the library. The placeholder suggests the provider’s default model.
Base URL	The API endpoint. Optional for cloud providers (defaults to the cloud endpoint) and required for local providers. The active and default URLs are shown beneath when relevant.
Timeout (ms)	How long to wait for a response, in milliseconds. Leave blank for the default — 60s for cloud providers, 5 min for local providers.
Test Connection	Pings the provider with the current key/URL/model and reports success or failure along with the response latency in milliseconds.
Save <provider>	Persists this provider’s API key (only if newly entered), model library, active model, base URL, and timeout — plus any pending default-provider change.

Caution

Keep your API keys secret. Treat them like passwords: never paste them into shared screenshots, chat messages, commits, or public links. Anyone with your key can spend on your account. Prefer supplying keys via environment variables on shared servers, rotate a key immediately if it is exposed, and use the eye toggle only when you need to verify what you typed.

Browser Sessions

The Browser Sessions tab lists every active browser session this Testver server is tracking and lets you close them. The list refreshes automatically every few seconds, and you can also click Refresh to update it on demand. Sessions come from three sources, each with its own colored badge:

Source	Badge	What it is
Server	Server (brand)	A server-side per-tab browser process — one per active /ai tab using server-side mode. Shows start time, process id (pid), and tab count.
Local	Local (green)	A local-browser connector running on a user’s own machine. Shows connect time, machine, OS, and tab count.
Test run	Test run (amber)	An active test-execution run (for example npx playwright test). Shows start time, framework, pid, and the running command.

Header and summary — the top card shows total counts and, for admins, controls to widen the view:

Element	What it does
Show all users / Showing all users	Admin toggle that switches between only your sessions and every user’s sessions.
Refresh	Reloads the session list immediately.
Summary stats	Shows totals — total sessions, number of users (when more than one), server-side count, and idle count.

Per-session row — each session shows its browser engine icon, source badge, browser type, owner (in all-users view), and an Idle indicator when inactive for over 5 minutes. The Close button on a row opens a confirmation dialog describing exactly what will be torn down (and warns if active tabs are using it) before terminating that browser process.

Bulk close — when sessions exist, a Bulk close card offers scoped shutdowns. Closing is graceful: the backend waits up to 5 seconds for in-flight actions to settle. Each scope prompts for confirmation first.

Bulk action	What it does
My sessions	Closes every browser session you currently own. Available to all users.
All server sessions	(Admin) Closes every server-side per-tab browser process to free server RAM; leaves local connectors and test runs alone.
Idle > 30m	(Admin) Closes every session idle for more than 30 minutes; active sessions are unaffected.
All sessions	(Admin) Closes every active session for every user, including local connectors and running tests. Use only during maintenance windows.

Caution

All sessions terminates every user’s browser mid-action. Anyone currently working will see their browser close. Reserve it for maintenance windows.

Common Tasks (How Do I…?)

I want to…	Do this
Fix how a single test runs	General → Run Commands → Single Test, edit the command (use placeholder tokens), click Test, then Save Run Commands.
Let AI figure out my run commands	General → Run Commands → Suggest with AI, review each suggestion, click Apply or Apply all, then Save Run Commands.
Add environment variables for test runs	General → Run Commands → Advanced Settings → Environment Variables → Add Variable, fill KEY and value, then Save Run Commands.
Run a build step before tests	General → Run Commands → Advanced Settings → Before Command, enter the command, then Save Run Commands.
Connect an AI provider	LLM, expand the provider, enter the API Key, add/select a model, click Save <provider> (use Test Connection to verify).
Add a model to a provider	LLM, expand the provider, type the model name in the add-model box, press Enter, star it to make it active, then Save.
Choose the default AI provider	LLM, click the star on the provider card you want as default — it switches immediately.
Use a local model (Ollama / LM Studio)	LLM, expand Ollama or LM Studio, set the Base URL, add the model name, then Save — no API key needed.
Switch between Web and Mobile	General → App Mode, click the other tile, and confirm in the dialog.
Switch light/dark theme	General, use the theme toggle — it applies instantly.
See which browsers Testver is running	Browser Sessions — the list auto-refreshes; click Refresh to update now.
Close a stuck browser	Browser Sessions, click Close on the row and confirm; or use Bulk close → My sessions.

Tips & Best Practices

• Use the Test button on each run command and Test Connection on each provider before relying on them — it catches typos and bad keys early.
• Prefer placeholder tokens ({file}, {test}, {posixFile}) over hard-coded paths so commands stay portable across machines.
• On Windows with WebdriverIO, use {posixFile} for the --spec argument to avoid backslash issues.
• Let Suggest with AI do the first draft of your run commands, then refine — it can reuse existing package.json scripts.
• Supply cloud API keys via environment variables on shared or CI servers so secrets never live in saved config; Settings will show the source as Env Var.
• Keep a small, focused model library per provider and star the one you actually use as the active model.
• For admins: prefer Idle > 30m over All sessions to reclaim resources without disrupting active users.
• Confirm Project Information matches the codebase you intend to test before configuring anything else.

Troubleshooting & FAQ

Problem	Likely cause & fix
A run command fails the Test	The binary or script can’t be resolved. Check the command, install missing dependencies, or Reset to the auto-detected default and re-test.
My edited command keeps reverting to the default	It only saves as a custom override if it differs from the auto-detected value. Make sure your edit is actually different, then click Save Run Commands.
Suggest with AI shows an error	AI couldn’t analyze the project. Use the Try in the AI Assistant instead link, or open Discuss in AI Assistant to work it out interactively.
Test Connection fails for a provider	Wrong or missing API key, wrong Base URL, or an unreachable local server. Verify the key (eye toggle), confirm the URL, and ensure local servers (Ollama / LM Studio) are running.
My API key field shows dots, not my key	The key is saved and masked for security. Click the field to clear it and type a new one; an env-var key shows a note naming the variable instead.
A provider stays Not Configured after saving	You may have saved without entering a key, or only changed a model. Re-open the card, enter the API Key, and Save again.
App Mode change hid a surface I needed	App Mode is workspace-wide. Switch back via General → App Mode; saved schedules and history remain on disk and reappear.
A browser won’t appear in Browser Sessions	Sessions only show when you launch a browser via /ai, register a Local Browser connector, or start a test run. Click Refresh.
I closed a session but it’s still listed	Bulk close is graceful and waits up to 5s for in-flight actions. Wait for the auto-refresh or click Refresh.
The theme didn’t change	The theme is stored locally in your browser; clearing site data resets it. Toggle it again from the General tab.

Related

• Local Browser Connector — token-based local browser routing.
• Connectors — external integrations.
• Admin → Users — user-level security.