MMMM Whitepaper

Solo Mode

In Solo mode, after equipping the Mic-Doll, users can select a song and independently complete the singing. Based on the singing evaluation, users can earn MESS. Users can only earn MESS when they publish their singing works, and each publication requires 1 point of Energy.
MMMM has a huge library of copyrighted lyrics and music, including song accompaniments, lyrics, and standard pitch lines used for singing evaluation. When users sing, the app will synchronize the accompaniment, lyrics, and standard pitch lines. Users can record singing with rhythm and pitch, and finally generate a complete song. MMMM also provides a tuning function.


After completing singing, users need to publish their singing records to earn MESS, and each publication requires 1 point of Energy. Energy has an upper limit, and every 6 hours, it will recover by 25%, but it cannot exceed the upper limit. At the beginning, the user's Energy limit is 2 points, and with more Mic-Dolls, the Energy limit will gradually increase.
The detailed table is as follows:
Mic-Doll Counts
Energy CAP

Singing Evaluation

Solo mode aims to allow users to enjoy singing. However, for the entire community, the better the singing works, the higher the value, and the more rewards they should receive. MMMM has developed a singing evaluation system based on AI technology, which can give a rating for the entire song based on the performance during singing:
  • For each sentence sung, AI will give a single score of 0-100, based on factors such as rhythm, pitch, emotion, and skill.
  • Based on the single score, Miss, Good, Great, and Perfect 4 levels of single-sentence evaluation will be given. The Accuracy attribute of Mic-Doll will affect the single-sentence evaluation. For example, if the single score is 70, a low Accuracy will only give a Good rating, while a high Accuracy will give a Great rating.
  • After singing a song, based on the cumulative single-sentence evaluation, a rating of C, B, A, S, SS, or SSS will be given for the entire song. The rating of the entire song will directly affect the amount of MESS earned.
We still believe that everyone can enjoy singing, so except for the lowest C-level, users can receive varying amounts of MESS for other ratings. The better they sing, the more they will receive. If users sing seriously, almost no one can only get a C-level. (In some cases, incorrect ratings may be given due to the material quality of the song pitch line. When this happens, please give us feedback in the App or on MMMM Discord, and we will continue to improve the quality of the library.)
In addition to singing evaluation, the amount of MESS obtained is also affected by the Mic-Doll Attributes and Mic-Doll Mode.

MESS Settlement

The calculation method for the amount of MESS earned each time when singing is as follows:
MESSPerSong=BaseValueWillpowerGrowthValue(RatingFactor+FeverFactor)AbilityFactorRandFactorMESSPerSong = BaseValue * Willpower^{GrowthValue}*(RatingFactor+FeverFactor)*AbilityFactor*RandFactor
  • Willpower: the Willpower of the Mic-Doll used by the user to sing the song, after Spirit attenuation calculation and Mic-Doll Ability bonus, the final effective Willpower value is calculated.
  • BaseValue & GrowthValue: are adjustable values set by the system and will be dynamically adjusted based on the balance of the entire economic system
  • GradeFactor: the rating coefficient of the song the user sang. The better the user sings, the higher the rating, and the user can see it on the settlement page after singing. The Accuracy attribute of Mic-Doll will make it easier for users to get high ratings.
  • FeverFactor: if the user successfully enters the Fever state during singing, there will be a bonus coefficient, which the user can see on the settlement page after singing.
  • AbilityFactor: if the user's Main Mic-Doll reaches level 10 and the Mode of the song sung matches the Mic-Doll, the Main Ability will be activated. If the user selects a valid Support Mic-Doll before singing, the Support Ability will be activated during settlement.
  • RandFactor: a random value with a mean of 1.
Overall, the amount of MESS earned by users each time when singing is mainly related to the Mic-Doll's attributes, singing evaluation, and whether the user enters the Fever state.

Mic-Doll Spirit Cost and Recover

After successfully publishing singing works, the Spirit of the Mic-Doll used will be cost and need to be recovered. The higher the level of the Mic-Doll, the more MESS is needed to recover.
Mic-Doll's Vitality attribute affects the rate at which Mic-Doll Spirit is consumed. The higher the Vitality, the slower Spirit is consumed. Low Spirit affects Mic-Doll's attributes.
  • When Spirit ≤ 50%, Willpower is reduced to 80%.
  • When Spirit ≤ 20%, Willpower is reduced to 10%.
  • When Spirit = 0%, Mic-Doll cannot be used for singing.
The calculation method of Mic-Doll Spirit Cost is:
SpiritCost=BaseValueVitalityGrowthValue100SpiritCost= BaseValue * Vitality^{GrowthValue} * 100
  • BaseValue: A system-set constant that releases the base factor of Spirit consumed in a song.
  • GrowthValue: A constant set by the system, which is a decimal number between [-1, 0], meaning that the higher the Vitality, the less Spirit is consumed.
The calculation method of Mic-Doll Spirit Recovery is:
RecoverCost=WillpowerGradeFactorBaseCostlevel/30SpiritRecoverCost = WillpowerGradeFactor *BaseCost^{level/30}*Spirit
  • WillpowerGradeFactor: willpower growth rate factor of Mic-Doll. The higher the growth rate grade, the more MESS will be needed for recovery.
  • BaseCost: the constant set by the system, the base factor of the MESS required for recovery.
  • Level: the level of Mic-Doll.
  • Spirit: the Spirit that Mic-Doll has lost.

Loot Box Dropping

After users complete singing and successfully publish their songs, they have a chance to get 1 Loot Box. Loot Boxes can contain Mic-Doll upgrade materials, Mic-Doll shards, and MESS. Each user has a total of 4 Loot Box slots, which means that users can only hold up to 4 unopened Loot Boxes at the same time. After getting a Loot Box, users can start unlocking it for free. After waiting for a period of time, the Loot Box will be unlocked and users can open it for free to get the rewards inside.

Loot Box Qualities

Loot Boxes are available in Common, Uncommon, Rare, Epic, and Legendary qualities. The higher the quality, the longer the time required to unlock the box, and the more rewards will be obtained. The quality of the dropped Loot Box is related to the probability and the Level of the Mic-Doll used during singing. The higher the Level, the better the quality of the dropped Loot Box.

Critical Loot Box

There is a certain probability that the Loot Box will be a Critical Loot Box, which will give double rewards when opened. The probability is affected by the Inspiration attribute of the Mic-Doll used during singing:
LootBoxCriticalChance=max(BaseValueInspirationGrowthValue/100,BaseChance)LootBoxCriticalChance=max(BaseValue *Inspiration^{GrowthValue}/100, BaseChance)
  • Inspiration: the Inspiration of the Mic-Doll used by the user to sing the song, after Ability bonus, the final effective Inspiration value is calculated.
  • BaseValue & GrowthValue: are adjustable values set by the system and will be dynamically adjusted based on the balance of the entire economic system.
  • BaseChance: the minimum critical probability.


We hope to reward real people who sing sincerely with MESS, rather than cheaters or robots. In order to protect the rights of community users, we have spent a lot of effort designing the anti-cheating system M-ACE, which can accurately identify non-human singing (such as playing the original recording) and cheating behaviors of using multiple devices to sing at the same time through face recognition, audio data, and abnormal behavior detection, combined with machine learning and a large amount of training data. Cheating behavior cannot obtain any Token rewards, and in severe cases, it may even lead to account penalties.
We use a variety of methods for identification and detection:
  • Real-time monitoring of pitch/rhythm/volume/lyric accuracy data for singing parts of the song
  • Sensor data for recording audio and video on the device
  • Common face and voice security detection
  • Abnormal recognition based on vocal and tuning principles
  • Cheating detection and processing based on machine learning and big data
M-ACE provides security for users' singing and recording and will remind or penalize users when it detects abnormalities. For most users, you will see anti-cheating prompts after singing is completed and obtain fair rewards based on the detection results, or be required to accept further verification and processing.
We welcome users to follow basic behavioral norms, sing and communicate in a friendly and active atmosphere. The following behaviors can help maintain a good usage status or improve the evaluation of past behaviors by M-ACE:
  • Maintain good behavior records and protect your wallet mnemonic and account information to avoid theft or fraud.
  • Sing genuinely and naturally, adjust devices and singing status as necessary according to the prompts in the game.
  • Treat the release of singing content carefully and reduce interference and misleading to the audience when they enjoy music.
  • Actively participate in community activities and provide feedback. Communicate with us in a timely manner when encountering abnormal situations.
Last modified 13d ago