testosaurusdocs
Browse docs
On this page

Dashboard & data

Testo celebrating

The dashboard is where reports land, get triaged, and turn into decisions. It lives at /dashboard and signs in with Better Auth.

Projects & tokens

Every project has public app tokens (the apiKey you pass to init). Reports arrive in the project the token belongs to. Copy, rename or revoke a token under settings → install & domains → App tokens: copy token copies it, rename… is in the token's more menu, revoke asks you to confirm, and new token adds another. A revoked token stops working immediately, and tokens stay listed and copyable otherwise. The hosts allowed to send reports are listed one per entry in the Allowed addresses card on the same page.

Views

The top bar holds the app switcher (Switch app), four tabs, search and commands, and the account menu. The settings sections grouped under this app apply to the app chosen in the switcher; every other section covers the whole workspace.

  • inbox — shared issues with report counts, status, severity, category and text filters. Open an issue for its shared AI summary and every original report, or switch show from issues to reports to browse individual captures. Review possible matches and correct grouping with merge or split.
  • insights — aggregates over time so patterns survive individual reports. When some reports couldn't be analyzed, it says how many and links to your reports; it does not list them.
  • overview — an app's setup steps, recent numbers, top issues, agent activity and where its issues are sent.
  • settings — eleven sections in three groups, listed below.

Press ⌘ K (ctrl K elsewhere) for the command palette, or / to filter the inbox. ? shows every keyboard shortcut, and g followed by i, n, o, s, a or b jumps to inbox, insights, overview, settings, agent access or billing & usage. The status bar shows whether the dashboard is live, whether your coding agent is connected and when it was last active, how many reports your plan has used and how many apps you have.

Settings has three groups. Members see every section; pages only the workspace owner can change say so.

  • this app
    • general — rename, archive or delete the app, or delete all of its reports.
    • install & domains — the snippet, Allowed addresses, App tokens and Bot protection.
    • widget — launcher, appearance, wording and capture choices.
    • ai analysis — the analysis switch, models by task, custom categories and fix setups.
    • alerts — Notifications and the Webhook card.
    • data — Storage and Export.
  • account
    • agent access — Connect a coding agent and your Access keys.
    • integrations — connect GitHub and Linear for the workspace; each project chooses its repository or team under Send issues to.
    • team — invitations, members and seats.
    • billing & usage — the meters, your plan, credit, usage alerts and invoices.
  • you
    • account & security — email, password, devices and account deletion.

Coding agents get their access keys under settings → agent access → Access keys, and the Connect a coding agent card on that page hands an agent a one-time setup prompt. See Connect your coding agent.

Limits & spending policies

PlanProjectsSubmissions / monthIncluded AI analysis
Free1100€1.00 / mo
Pro1010,000€3.00 / mo

For larger requirements, contact us to agree resource allowances, included AI analysis and support terms.

Each resource stops at its limit unless you turn on extra usage for it, where your plan allows paid overage, and you can set alert thresholds. Submission and AI allowances reset monthly; reaching them stops new submissions or new AI processing respectively. Storage and seats stay occupied across months: free up storage to make room for new uploads, or remove members to make room for new invitations. An upgrade or eligible paid overage can add capacity. See Pricing.

Overage is opt-in

Each metered resource on settings → billing & usage either stops at its included limit (the default) or continues as extra usage. On Pro and Enterprise the page has one line per resource, such as extra usage is off · turn on; turning it on shows the published rate and asks you to confirm, and turn off reverses it. Overage needs two yeses — a plan that supports it (Pro or Enterprise), and your opt-in for that specific resource. Missing either, usage stops at the included limit and nothing bills. With both, usage past the included allowance continues at the published rate and is drawn from your prepaid balance at period close. Every resource also has an alert threshold: under usage alerts, email me at takes off or a share from 10% to 100% of the allowance. Cross it and Testosaurus emails you, so surprises arrive as mail, not as charges.

AI usage follows the same opt-in but meters input and output tokens at the current catalog base-token rates plus 5%, against your plan's included AI analysis. The catalog converts gateway rates to EUR and does not reconcile provider invoices or cached-token discounts. Once the included AI analysis is used up and you've opted in, AI charges draw from your prepaid balance in real time. When the included AI analysis and your credit run out, new distillation stops. Feedback remains available, subject to submission and storage limits; a top-up queues the reports skipped for lack of credit again. The daily AI spend cap can also stop processing until the next UTC day. Free has no overage, so its €1.00 of included AI analysis stops processing when it is used up. Alert thresholds apply to the included AI analysis. The AI analysis by model fold on the same page itemizes AI calls, tokens and metered euros per model.

Automatic top-ups and monthly credit

Under settings → billing & usage → billing details and invoices, workspace owners on paid plans can authorize a saved refill payment method and enable automatic top-ups. Choose when the balance is low, how much to add and a monthly automatic-purchase limit. Saving a payment method does not enable charges. The limit follows UTC calendar months and includes pending purchases; turn off refills at any time, allowing payments already processing to finish.

The Monthly credit block there adds a fixed amount to your existing monthly Pro invoice, starting at the next renewal. Credit is granted after payment and rolls over in the same prepaid balance as other top-ups. Change or stop the credit line while keeping Pro, and buy one-off credit with add credit on the same page. Failed payments grant no credit, and funding does not change your resource overage policies. See Pricing & limits.

Export

Every plan can export a project as JSON, whenever you want: no lock-in, no ransom, no waiting on support. The export is a portable copy of your reports, not a backup.

The export data button lives in settings → data → Export and downloads one JSON file per project: its settings, every submission with its raw text, context and distilled fields, and a manifest of the stored screenshots and replays. It does not include the screenshot and replay files themselves, shared-issue groupings and their history, coding-run history, tokens or the webhook secret. The same download is available as GET /api/projects/:id/export — see the API reference for the exact shape.

Turnstile modes

Testosaurus protects every project with a shared Cloudflare Turnstile widget for the mode you choose, so there is nothing to provision. Invisible runs silently; Managed can show an interactive challenge when needed. New projects default to invisible. Choose the mode under settings → install & domains → Bot protection, in the advanced bot protection fold. If you bring your own Turnstile keys, manage their mode in your Cloudflare dashboard and list testosaurus.dev in that widget's hostnames, because the challenge runs there.

AI settings

Distillation control lives in settings → ai analysis. The Analysis card holds the switch that turns AI analysis on or off for the app; with it off, new reports are kept as they arrive, without an AI summary. The AI models by task card assigns models to roles instead of picking one provider, and each role meters tokens at its model's catalog rates plus 5%:

  • Summarize reports (the Default) — matches differently worded reports and builds or refreshes shared issue summaries. Exact known repeats reuse the existing analysis. It also routes new issue analysis to a deeper analysis role when appropriate. It can be left on Testosaurus default.
  • Read screenshots (Vision) — optional. Takes the Default's place when a report carries a screenshot; the list only offers vision-capable models.
  • custom categories — in the advanced fold, add a category role: you name it, pick the model and write its instructions, for example "Bug investigator" on a reasoning model or "Idea synthesizer" on a big generalist. The semantic router and the category rules dispatch deeper analysis to these roles.
  • testosaurus-run fixes — in the other advanced fold, each fix setup (a coding role) has a model, instructions, base branch, allowed paths, test commands, an optional dependency setup command, and budgets per run and per month. Install the reviewed GitHub Actions workflow from the same fold, then use Propose fix on an analyzed report. Optional automatic categories start a fix setup after matching reports arrive. Coding roles produce tested draft PRs for human review.

auto-assign, inside the custom categories fold, fills the default model, a screenshot model and a deeper-analysis role from the model catalog in one click, preferring models that support structured output (and vision, for screenshots). Everything stays editable afterwards, and nothing is saved until you press save. Under Send each category to, the four categories (bug, confusing, idea, other) can each be pinned to one of your category roles, so, say, every bug report is deepened by the bug role's model. A rule beats the router when both would fire.

The model lists show the chat models Testosaurus offers through the Vercel AI Gateway, grouped by provider. Under them the card says about how much one report costs with the models you chose and the most a very long report can cost, so the price is visible before you save. For the workspace owner it also shows this month's AI analysis used against the included AI analysis, with a link to billing & usage.

Coding reserves the role's maximum run budget from prepaid credit and meters actual model tokens at the same catalog rates plus 5%; unused reserved credit is returned once. Your GitHub Actions account runs the job, with no additional Testosaurus execution fee. Run progress, test results, spending and the draft-PR link appear on the report. Only the workspace owner can start or cancel runs; publication already in progress completes its draft handoff. See Reports & distillation.

Invite your team

Under settings → team you can invite teammates to your workspace by email. Each invite has a link you can copy and send yourself if the email does not arrive. Invites expire after 14 days; re-invite or revoke an expired one at any time. Accepting an invite switches the teammate's dashboard to your workspace, where they read and triage reports; only you change settings, integrations, tokens, billing and storage. Teammates can leave the workspace from settings → team → Leave workspace. A plan's included seats cover you plus your teammates. A pending invitation holds a seat but is never billed. Only accepted members count toward billing: every seat beyond your plan's included seats bills the published seat rate from your prepaid balance, and removing a member frees the seat immediately. You delete your own account under settings → account & security → Delete account.

Submission detail views have stable URLs, so "look at this one" is a link you can paste into Slack or an issue: /dashboard/?id=…

Something unclear?Let us know