Questions
Start with your question
Know what to expect, and where to ask
Find answers from choosing a mission to reading and sharing a result. The last part explains why we are building OperatorRank.
- Before starting
- Cost and preparation
- While working
- AI and interruptions
- After submission
- Results and Defense
- Access and organisations
- Who reads and decides
- About us
- What we are building
Before starting
Before starting: does it cost anything?
| Question | Answer |
|---|---|
| Do I need a card? | Nothing is charged today and no card is required. Missions still use credits. |
| How do I get free credits? | Verify with Google to get free credits when eligible. Pricing explains today’s amounts and expiry. |
| Where is this attempt’s cost? | Use Start for this attempt’s quote; Pricing explains funding. |
What should I prepare before choosing?
Read the brief before starting; you do not need to commit just to learn the task.
-
Does an account open every mission?
Availability, version, eligibility and funding still apply. An account alone does not satisfy every start condition.
-
Do I install mission tools?
Workspace work happens through the browser, with terminal execution in the mission environment. A larger screen helps with files.
-
Which language is used?
The site, Start and results support Thai and English. The workspace is in English; the brief supplies the mission’s content language.
While working
While working: how much can the AI do?
It can act, and you inspect what happened
The AI agent can read, edit, and run allowed tools within the attempt's model menu, budget, and time. Always inspect the outcome after directing it—AI usage volume isn't the pass rule.
Choose the relevant workspace controlDiagram - Direct an action
- AI conversation
- Inspect the outcome
- Files, diffs and commands
- Change direction
- Stop and inspect state
What if budget, connection or time runs short?
| Situation | Next step |
|---|---|
| AI request does not fit the budget | Manual tools keep their own gates. Use available files, terminal and Submit within the remaining time. |
| Disconnected or page closed | Use the displayed recovery link to return to the same attempt; see the mission guide for interruption rules. |
| Submission review opens | Return to work before confirming if needed; the mission guide covers the full flow. |
After submission
After submission: when is the result, and how do I read it?
How long does it take?
The result page shows an estimate where available. Queueing and mission size matter. You can return later; there is no fixed turnaround guarantee.
What if it is Not passed?
Find the piece below its bar and inspect the work. Another attempt needs mission availability, account eligibility and funding.
What if the result looks wrong?
Results are provisional and can be corrected through review. Report issues before the result page's deadline and track your request.
What is the Defense conversation?
An unscored discussion of the submitted work
On enabled missions, an AI reviewer asks from the supplied work record. Defense may be absent, optional or required; follow the attempt’s window. It preserves an explanation for authorised readers, without certifying identity or infallible understanding.
Read this mission’s Defense settingsDiagram - Availability
- Enabled per mission
- Requirement
- Attempt-specific mode
- Score
- Unscored
Access and organisations
Access and organisations: who can read my work?
| Question | Answer |
|---|---|
| Do recruiters see my practice? | Practice is not public. Shares, applications and group grants have separate scopes. |
| Does a summary share files? | A summary has its own scope; Who sees what lists the disclosed fields. |
| Can I request deletion? | Yes, through the account’s confirmation flow, with a tracking ID. Receipt is not completion. |
How do companies and classes use results?
For the full workflow, see For organisations. Pricing explains funding.
-
How does a company begin?
Set up the organisation and roles, then use eligible public links, rounds or job postings under current allowances.
-
Can a company set the pass bar?
Today, pieces and required levels come with the published mission. Reviewers do not invent a new pass rule.
-
Does AI make the hiring decision?
No. Results are one source of evidence for people to interpret in the company or school’s context.
About us
About us
We want real work with AI to be inspectable
Our starting question was simple: when AI can generate polished output, how do you see the thinking and judgment behind it? OperatorRank puts you in a simulated work scenario with a submitted snapshot, piece-by-piece grading, and the records available to inspect. We're starting with tech because software outcomes can be tested deterministically. That's our starting point within the wider world of work with AI, not its limit.
The principles and limits we build around
The corresponding pages explain each mechanism. The detail should be inspectable rather than resting on an advertising claim.
-
Say what was measured
An outcome belongs to the task. It is not an intelligence or personality score, or an employment guarantee.
-
Explain where data goes
Sharing and provider processing need explicit boundaries. Privacy does not mean nobody processes the information.
-
Make room for correction
Review and correction are standard. Keep evidence within its scope and keep a person in decisions about people.
A few more questions
Where can I ask another question?
Use Help in the app or the support address below. For an attempt-specific issue, report from its result page so the request carries the work context.
How do I suggest something about OperatorRank?
Send your suggestion and a concrete example to support@operatorrank.com.
Ready? Start with the brief
Sign up, choose work you want to try, and read the rules before starting.