Concepts
The words used throughout Cettara, and what each one means.
Organization
Your company. It owns projects, people, billing and one audit log. Somebody can belong to more than one.
Project
A workspace for one kind of testing — manual or automated — with its own cases, settings and people. Visible to the whole organization, or only to named members.
Test case
The test itself, not a document describing it. A manual case is steps and expected results; an automated case is steps and checkpoints.
Step
One action. Automated steps include going to a page, clicking, filling a field, attaching a file, waiting, calling your backend, asking the AI to do something in plain language, and branching.
Checkpoint
The moment an automated run stops and asks: is this right? You write the expectation; the AI judge answers with a verdict, its confidence, its reasoning and the evidence it used. One checkpoint costs one AI credit.
Block
A reusable set of steps — a sign-in, a checkout — written once and used by many cases. A block marked run-once runs a single time per run and every case shares the result.
Flow
An ordered group of cases that run together, each with its own values.
Milestone
A release. It gathers the cases that must pass, shows progress, and produces a report somebody signs.
Run
One execution. It has a status, a number, the results of every case in it, and the evidence.
Verdict
What the judge decided at a checkpoint: passed or failed, with confidence, reasoning and named evidence.
Evidence
What a run leaves behind: video, execution trace, screenshots, downloaded files. Kept for 30 days on Free, 90 on Team, and up to seven years on Business.
AI credit
The unit for anything that asks the model. A checkpoint is 1, a question about the screen is 1, a Do step is 3 however many tries it takes, and a checkpoint that reads a PDF is 2. Every plan includes an allowance each month, and more cost two cents each.