One prompt. Twenty seconds. Fair rules.
These rules describe the private-beta format. Ranked LP remains disabled until live-media integrity and judging quality pass the documented release gates.
Larpit is currently available only for private development testing. These pages describe the intended beta controls and will be reviewed for each launch country before public availability.
These operational drafts do not replace country-specific legal review.
1. Round format
- Two eligible users are matched into a private real-time room.
- Both users receive the same server-selected, pre-reviewed presentation prompt.
- Camera and microphone are required for the current test format. Microphone audio is live only and is not recorded or judged.
- The live challenge lasts 20 seconds after both users are connected and ready.
- Selected still frames near seconds 3, 8, 13, and 18 may be used for evidence and visual judging.
2. What determines the result
Each submitted entry is evaluated on four observable categories, up to 25 points each:
- Prompt completion: the required safe elements or actions are visibly completed.
- Sequence and timing: the sampled moments form a clear beginning, development, and finish.
- Composition and legibility: the showcased content is deliberately framed and understandable.
- Presentation coherence: the visible arrangement, reveal, and theme work together.
Identity, attractiveness, face or body, age, gender, race, health, disability, religion, sexuality, emotion, personality, confidence, wealth, status, popularity, brand, price, rarity, ownership, home quality, accent, language, followers, or existing rank.
3. Automated judging safeguards
The intended beta design uses balanced evidence, random participant labels, deterministic validation, image safety screening, a fixed versioned rubric, and a structured result. A close or uncertain result is not forced into a winner. Audio, usernames, follower counts, prior rank, and profile history are not sent to the visual judge.
4. Win, draw, and void
- A confirmed win requires sufficient safe evidence and a result outside the configured draw band.
- A draw is a completed valid match with no clear winner. Both players receive 0 LP.
- A void means evidence, safety, connectivity, or integrity requirements were not met. It does not count as a competitive result and both players receive 0 LP.
- During the test season, intended win/loss movement is +24 LP / −18 LP, with LP floored at zero. These values may change before launch.
5. Disconnects and leaving
- Leaving before the live round begins cancels the attempt without LP.
- Choosing Leave after the round begins is a forfeit.
- An unexpected one-sided loss gets a five-second reconnect grace.
- A forfeit is applied only when server/provider evidence shows the opponent remained and the match was active for more than five seconds.
- Early, mutual, provider-wide, or ambiguous outages are void with 0 LP.
- Three abnormal disconnect or evidence failures in 24 hours may trigger a 24-hour ranked cooldown, subject to appeal.
6. Limits, rewards, and integrity
Initial private testing is limited to three judged matches per user per day, one active match at a time, and at least 60 seconds between starts. Public defaults may increase after safety and cost testing. Larpit+ will not sell extra ranked attempts, LP, preferred judging, or match outcomes. There are no cash, redeemable, or wagered prizes.
7. Appeals
Either participant may appeal within 24 hours for wrong evidence, evidence mismatch, prohibited-factor use, moderation error, or result-mapping error. An appeal is a human integrity review, not a request to reroll the AI. Evidence placed on appeal hold is normally deleted when review ends and no later than seven days unless a documented legal or urgent safety hold applies. Invited testers must open the appeal through the in-app flow when available, or use the contact channel provided with their invitation. Public email routing is not active yet.