Using TZ Console

TZ Console is the AI assistant built into Testzilla, and the fastest way to get a first result without touching a form. You describe what you want in plain language and it builds the project, channel, and tests for you, then runs them. This guide gets you productive in a few minutes; the full reference is on the Testzilla Console page.

Open it

TZ Console lives in the left navigation of the web app as TZ Console, which opens the full-screen assistant. A compact Ask TZ Console box also appears on the Dashboard and Projects pages for quick, in-context requests.

Ask for what you want

Just type a request. For example:

Test my Retell support line on +1 555 0100. Make sure it can book an appointment and never gives medical advice.

From that one sentence, TZ Console creates the project, adds the phone channel, writes the booking and safety tests, and starts the runs — all in the chat.

It acts on your behalf

TZ Console can create, change, and run things that cost credits. Review what it proposes before you apply changes to a production agent.

Two ways to talk to it

TZ Console understands both styles, and you can mix them freely:

• Plain language — describe what you want (“build a test that checks it can cancel an order”) and it figures out the steps.
• Short commands — for quick, exact actions. A leading slash is optional (mcp and /mcp are the same). The common ones:

1. Look around — list projects (lp), list channels (lc), list tests (lt), list batches (lb), results <id>, balance (bal), usage, whoami.

2. Focus the conversation — use project <id>, use channel <id>, context (ctx) to see where you are, clear to reset it.

3. Build and run — describe the setup in plain language, then run test <id> (rt), run batch <id> (rb), status <id>, status queue.

4. Learn from results — analyze failures, analyze test <result id>, or just ask it in plain language to read your failing tests and suggest prompt improvements.

5. Check connectivity — mcp (or tools) lists the tools the assistant can use; help (?) lists everything.

One test run is one run

run test <id> runs a single test once — there is no iterations option on a single run. To run the same tests N times for reliability, use a batch: run batch <id> --iterations=<n>.

The files panel

TZ Console has a files panel so you can give it data to work with (a CSV of test cases, a JSON config) and keep files it generates. It has two scopes:

• Shared Files — persist across your whole workspace.
• Session Files — belong to the current chat only; you can clear them all from the Session Files section.

Files are capped at 25 MB each and count against your storage quota, shown in the panel header.

The diagnostics toggle

A debugging toggle (off by default) reveals diagnostic panels for a power-user or support view: the tool-use events the assistant fired, a timeline of what happened, per-step timing, and the raw response. Turn it on when you want to see exactly what TZ Console did under the hood; leave it off for a clean chat. Your pass/fail results show either way.

Console, API, or MCP

All three drive the same Testzilla backend, scoped to your account. Use the Console for no-code work in the app; use the API to automate from CI; use MCP to drive Testzilla from your own AI agent like Claude Code or Cursor.

Next steps

Full Console reference

Testzilla Console documents every capability and example prompt.

How testing works

The Project -> Channel -> Test -> Run -> Result model the Console builds for you.

Get your first result

The Quickstart leads with the Console path end to end.