Describe your own ethical situation in normal language.
Quick Start
Studio helps you ask one main question: is this choice both ethically sound and practically wise? You can write your own situation, start from a sample, compare saved work, or run several cases at once.
-
Open
/studioand sign in. If you have admin access, the Admin link appears in the top bar. - Write the case you want to review. Include the choice, people affected, setting, stakes, and any risk you already see.
- Add context if it matters. Choose Custom to set the people affected, setting, decision type, and stakes.
- Select Evaluate. A custom evaluation uses credits. Sample cases and saved reviews do not spend credits.
- Review, save, export, or improve the result. Use the follow-up tools when you need a stronger answer or a safer rewrite.
Workspace
The Studio screen has two main areas. The left side is where you write and evaluate a case. The right side is where you can browse sample cases, saved evaluations, and batch tools.
Pick an example from the Jymni Guide and run a similar review.
Reopen, compare, or delete past evaluations.
Credit note: Custom evaluations use credits. The screen shows credit cost before higher-cost actions run.
Run an Evaluation
A good case gives the evaluator enough facts to make a useful call. You do not need perfect wording. Short, clear facts work best.
Live Evaluation Spec: Studio uses the live Evaluation Spec chosen by an admin. The spec tells Studio what to score, what facts to check, what tone to use, and how to shape new results.
Write the use case
- Choice: What you are considering saying or doing.
- People affected: Who may benefit, be burdened, or lose a meaningful choice.
- Setting: Where it is happening, such as home, work, school, online, healthcare, or the community.
- Facts: What you know, what you assume, and what remains uncertain.
- Concern: Any possible pressure, privacy issue, unclear term, or harm.
Use the optional guide
Select Use guide if you want help turning rough notes into a clearer case brief. The guide does not submit an evaluation by itself.
Add optional context
Choose Custom when the relationship, setting, decision type, or stakes should change how cautious the review is. This is useful for vulnerable people, health or money topics, unequal power, privacy, safety, or high-pressure choices.
Read Results
The result starts with a clear classification, scores, and a short reason. Then it explains the practical effects, ethical issues, risk signals, missing facts, and possible improvements.
Admins may change which Evaluation Spec is live. Saved evaluations keep the result they already have. New evaluations use the spec that is live when they run.
| Result area | What it means | What to do next |
|---|---|---|
| Classification | The main label for the case: Aligned, Self-Serving, Well-Meant, or Reckless. | Use it as the quick read, then check the reasons before deciding. |
| Ethical impact | How fair, honest, careful, and respectful the choice appears. | Check facts, consent, privacy, safety, and respect when this score is weak. |
| Practical wisdom | How workable, proportionate, sustainable, and repairable the choice is. | Make the value clearer without adding pressure or confusion. |
| Risk Delta | The main drivers that made the case more risky. | Open it when you need to know what to fix first. |
| Actionable improvements | Specific edits or guardrails. | Use these as a practical next-step checklist. |
Useful result actions
- Save: Store the result so you can reopen it later.
- Ask Follow-up: Ask a focused question about the current result.
- Strengthen This: Create a safer, clearer version of the choice or conversation.
- Counterargument Lens: See how other affected people, a fair critic, or a trusted advisor might challenge the reasoning.
- Export PDF: Print or save the result as a PDF for review records.
Improve Results with AI Tools
After Studio finishes an evaluation, extra action buttons appear below the result. These tools help you question the result, improve the choice, and test a better approach before acting.
| Tool | What it does | Use it when |
|---|---|---|
| Counterargument Lens | Looks at the choice from another perspective. It shows how affected people, a fair critic, or a trusted advisor might object. | You want to find weak assumptions, unfair pressure, missing consent, or ways the choice could harm someone. |
| Run Lens Again | Runs Counterargument Lens again on the same evaluation. | You changed the case, want a second pass, or need more possible objections before approval. |
| Strengthen This | Creates a better version of the choice while keeping a legitimate personal goal. | The evaluation says the goal has merit, but the approach needs clearer facts, calmer communication, better boundaries, or stronger safeguards. |
| Ask Follow-up | Lets you ask a specific question about the completed evaluation. | You need a clearer answer, a next-step checklist, a risk explanation, or help deciding what to address first. |
| Ask Another | Appears after you ask one follow-up question. It lets you keep asking about the same evaluation. | You have a second question or want to build a short reflection thread before revising the plan. |
How to use Counterargument Lens
- Run a normal evaluation first. Counterargument Lens works from the completed result on screen.
- Select Counterargument Lens. Studio lists objections, weak spots, likely misunderstandings, and trust risks.
- Review the suggested revision. If Studio provides a revised version, compare it with your original approach.
- Select Run Lens Again when needed. Use another pass if the case changed or you want more objections before approval.
How to use Strengthen This
- Run a normal evaluation first. Use the result to see what needs work.
- Select Strengthen This. Studio reframes the approach to be clearer, fairer, and more workable.
- Read what changed. Look for better facts, clearer consent, lower pressure, fairer burdens, and a realistic repair path.
- Evaluate the rewrite if offered. This checks whether the new version improves the score and risk signals.
How to use Ask Follow-up and Ask Another
- Select Ask Follow-up. Ask one focused question about the current evaluation.
- Ask a clear question. Good examples are “What should I address first?” or “What facts should I verify before acting?”
- Review the answer. Use the answer to decide whether to revise, save, or re-evaluate the case.
- Select Ask Another for the next question. Ask Another keeps the follow-up tied to the same evaluation, so the review stays organized.
Credit note: These tools may use credits. Studio labels higher-cost actions before they run.
Sample Cases
Sample cases help you learn the Jymni Guide before you write your own case. Choose a quadrant, review the list, then select a case that is close to your situation.
| Quadrant | Plain meaning | When to use it |
|---|---|---|
| Aligned | Ethical and effective. | Learn what ethical and practically wise choices look like. |
| Self-Serving | Effective, but ethically risky. | Spot pressure, deception, or self-serving rationalizations. |
| Well-Meant | Ethical, but not very effective. | Make a well-intentioned choice more realistic or useful. |
| Reckless | Weak on both ethics and effectiveness. | Learn what to avoid. |
Saved Evaluations
Saved evaluations let you keep a record of past reviews. You can reopen a case, sort the list, compare two saved results, or delete items you no longer need.
- Open Saved evaluations. Use the tab on the right side of Studio.
- Choose a saved case. The saved result opens so you can review it again.
- Compare two evaluations. Pick two saved items and select Compare evaluations to see score and context changes.
- Delete only when you are sure. Deleted saved evaluations cannot be restored from Studio.
Batch Evaluation
Batch evaluation is for checking several short cases at once. Paste one case per line, or separate longer cases with blank lines.
Before running a batch: Each completed case uses credits. Keep each case short and specific. Studio runs up to 20 cases from one durable batch; failed items are refunded.
- Select Batch evaluation. The dialog opens with a large case field.
- Paste your cases. Use one line per case for short cases. Use blank lines between longer cases.
- Check the estimate. Make sure the case count and credit total look right.
- Select Run evaluations. The server keeps working if you close the dialog or refresh. Finished results are saved automatically and appear in Saved evaluations.
Settings
Select Settings in the top bar to manage your account. The available panels depend on your role.
See personal or active-workspace balance and usage, buy a configured credit pack, open payment-provider receipts, or submit a structured trial/support request.
Create and select workspaces, invite members within the seat limit, manage roles and policy guidance, pool owner credits, record review decisions, and export workspace audit history.
Add or remove device sign-in methods for faster access.
Change your password when your organization requires it.
See how many credits each action costs.
Review terms and privacy information.
End your session on the current device.
Buy or request credits
The Credits panel shows checkout first. When payment plans are configured, choose a pack to continue to Stripe-hosted checkout. Completed purchases appear in purchase history, with a receipt link when Stripe supplies one. If checkout is not configured, Studio says so without hiding the available support path.
When your personal balance reaches the low-credit threshold, the top-bar credit control changes state and Settings shows a refill notice. View credit packs moves keyboard focus and the visible panel to checkout so you can act before a paid workflow is blocked.
Open Request trial or support credits only for an evaluation problem, beta or sales trial, nonprofit or educational use, or another reviewed exception. Choose a reason, amount, and intended use. Requests are capped, only one can remain pending, and most reasons require a low balance and observe a cooldown. After a recent paid purchase, Studio limits manual requests to evaluation-error, nonprofit/education, and sales-assisted-trial exceptions and explains unavailable choices before submission.
Invite and govern a team
Workspace owners and admins can invite an email address and assign an admin, reviewer, or member role. The invitation reserves a seat until it is accepted, revoked, or expires. If email delivery is unavailable, Studio shows a secure link for the inviter to copy to the intended person. The invitee must log in with the invited address.
The workspace owner can enable an owner-funded credit pool. Workspace AI actions then draw from the owner's balance while each member's personal work remains separate. Owners and admins can also download a CSV audit history of membership, invitation, policy, sharing, comment, review, and billing-control changes. Audit metadata intentionally does not duplicate prompt or comment text.
Tips and Troubleshooting
The Evaluate button is disabled.
Add more detail to the situation. Include what you might do, who is affected, and what each person is being asked to accept.
I ran out of credits.
Open Settings and go to Credits. Buy a configured credit pack, or open the secondary trial/support request when your situation qualifies for manual review. If the active workspace uses an owner-funded pool, ask the workspace owner to check that balance.
My workspace invitation did not arrive.
Ask the workspace owner or admin to confirm the address and pending invitation. They can revoke and resend it. When email delivery is unavailable, Studio gives the inviter a secure link to send directly. Log in with the exact invited email address.
The result asks for missing facts.
Add the missing facts and run a new evaluation. Common gaps include consent, relationship history, time pressure, alternatives, likely harms, and whether the choice can be reversed.
I need to share the result.
Use Share to create a read-only link. Choose public, signed-in, or workspace-only access and a 7- to 365-day expiration. You can inspect view counts, refresh the snapshot, or revoke access. Export PDF remains useful for a fixed review packet.
What does confidence mean?
Confidence describes how strongly the available case facts support the result. Check the Evidence & Confidence section for verification status, claims to verify, citations, and alternative ethical lenses. Every result still requires qualified human review.