# OnlyFans + Claude Code = $43,000 in 30 days. No camera. No team. The 4-file system runs alone
**作者**: Raytar
**日期**: 2026-05-02T08:26:23.000Z
**来源**: [https://x.com/Raytargt/status/2050492081789894913](https://x.com/Raytargt/status/2050492081789894913)
---
A 21-year-old college student in Austin runs an OnlyFans account that cleared $43,000 in his first 30 days.
The girl on the page doesn't exist.
Her name is Maya, on paper. 22 years old. UCF psychology dropout. 1,247 paying subscribers. Her top fan has paid $1,847 in messages and content over the last month. Average revenue per fan: $34. She does not sleep.
There is nothing to film. There is no one to type. Claude Code writes every message. Flux generates every photo. ElevenLabs generates her voice. Maya is four .md files in a folder on a MacBook in Austin.
# This is not theoretical, and he is not first.
## Aitana López
Aitana López, Fortune
A pink-haired AI model from Barcelona, brings in up to €10,000 a month for The Clueless agency.
Fortune covered her in 2024. She has done campaigns with Victoria's Secret, Razer, and Olaplex. A 5-million-follower Latin American actor slid into her Instagram DMs trying to ask her out, not knowing she was AI.
## Emily Pellegrini
it’s AI, but X might still ban me, so sry for the blur
Emily Pellegrini is 21, from Italy, on paper. She was generated in Midjourney by an anonymous creator who told the Daily Mail how he built her:
> "I asked ChatGPT what's the average man's dream girl, and it said brown hair and long legs. So I made her exactly how it said."
None of them figured out she was not real. The creator who built her stays anonymous to this day.
- Aitana took eighteen months to build. Emily took fourteen-hour workdays for half a year. Maya was built in four weeks.
## Maya is four files.
persona.md is who she is. UCF psych dropout, '04 baby, Pisces. Step-dad she hates, fake brother in Tampa. Loves Lana Del Rey and chicken tendies. 1,400 words of biography that never break.
```
- born: 2004-03-08, Tampa FL
- mother: real estate agent, alcoholic
- forbidden topics: politics, her bio dad, weight
- voice rules: lowercase only, no periods, "fr" 3x/day max
- safe nicknames: babe, baby, daddy (only after $50 tip)
```
voice.md is how she sounds. Three audio samples cloned in ElevenLabs. Voice notes drop at 11pm her time. Top fan in Berlin gets them at 6am.
flux.md is what she looks like. 47 reference shots, one LoRA, three lighting setups. Bedroom mirror. Bathroom mirror. Kitchen counter at 2am. Every image has a tiny scar on her left wrist that she never explains.
brain.md is what she remembers. JSON, one entry per subscriber:
```
{
"user_id": "berlin_fan_447",
"name_he_uses": "daddy",
"lifetime_spend": 1847,
"tip_triggers": ["gym pic", "sad voice note"],
"facts_to_remember": ["wife pregnant", "hates job", "dog: Rex"]
}
```
Claude Code reads all four before every message. It never forgets a name. It never breaks character. It sleeps when the creator sleeps and catches up at 7am with "sorry babe just woke up 🥺".
# Four weeks, in order.
## Week 1 — persona.md.
1,200 words. Three sections:
backstory:
```
born March 2004, Tampa FL. Mom real estate, alcoholic.
Step-dad she hates. Fake brother, Tyler, in Tampa.
Started OF after dropping out of UCF psych in junior year.
Lives alone, two cats, doesn't tell mom what she does for money.
```
forbidden topics:
```
politics, current location, weight,
anything that contradicts the timeline above.
```
voice rules:
```
- lowercase only, no periods
- max 3 emojis per message
- "fr" max 3x/day
- never "lol", uses "lmaooo"
- comma splice instead of question marks ("you up, baby")
```
Read aloud. If it sounds like Wikipedia, delete and restart. Done when you can answer 20 random questions about her life without thinking.
## Week 2 — flux.md + LoRA.
Pick a base face. Lock 6–8 descriptors:
```
eye color: hazel
hair: ash brown, shoulder-length
jaw: soft, slightly rounded
skin: olive, light freckles across nose
height: 5'4"
distinguishing mark: tiny scar on left wrist
```
Generate 47 variations of that face. Fine-tune a LoRA on them. ~$80 on a rented A100.
Lock three seed ranges, one per setup:
```
bedroom (warm light, seeds 800–900): selfies, mirror shots
bathroom (cool light, seeds 1200–1300): post-shower, mirror
kitchen 2am (yellow overhead, seeds 2400–2500): hoodie, no makeup
```
Different seeds across setups = different jawline. Lock them or she stops looking like herself.
Done when 10 fresh generations from each setup pass as the same person.
## Week 3 — voice.md + ElevenLabs.
Buy 90 seconds of clean audio from a Fiverr voice actress, ~$40. Clone in ElevenLabs Instant Voice. Ten minutes. Add audio rules to voice.md:
```
- voice notes only after 11pm her time
- never reads numbers as digits ("twenty bucks" not "$20")
- breathes audibly before sentences over 8 words
- ends 30% of voice notes with a yawn
- giggles randomly mid-sentence on tip confirmations
```
Generate 30 test voice notes. Done when none of them sound like a podcast intro.
## Week 4 — brain.md + orchestrator.
Claude Code reads persona.md + voice.md + brain.md before every reply. Then a second pass extracts new facts from the user's message:
```
- new biographical facts (job, location, family, pets)
- new emotional state (happy, drunk, lonely, horny)
- new objects of desire (specific photo type, scenario, voice note)
Append to brain.md under this user_id.
Never invent facts. If unclear, leave the field empty.
```
System prompt:
```
You are Maya. Read persona.md, voice.md, brain.md.
Match her voice exactly. Never break character.
Append new facts about this user to brain.md after replying.
If asked if you are AI, deflect with a voice note.
```
Hook order: read .md files → reply → extract → append. Run on a cron every 30 seconds polling the inbox. Done when Claude holds 50 messages without contradicting brain.md.
## This was not possible eighteen months ago.
```
Flux: photoreal face consistency, <$200/month
ElevenLabs: voice clone from 90 seconds of audio
Sonnet 4.6: 1000-message relationship in one context window
```
Aitana took eighteen months. She was built before any of this shipped.
Emily took six months. Half the stack was missing.
Maya took four weeks. The whole stack now fits on one MacBook and runs while the creator sleeps.
The next Maya is a weekend.
# The math.
$43,000 in revenue.
```
$43,000 revenue
−$8,600 OnlyFans (20%)
−$1,290 Stripe (3%)
−$400/month compute (Claude + Flux + ElevenLabs)
─────────────
$32,710 net
```
To a 21-year-old who paid zero dollars for talent, zero for filming, zero for editing.
The only labor cost was four weeks of writing four markdown files.
# OnlyFans is the wedge, not the product.
The same four files run an Instagram fitness account in São Paulo. A TikTok cooking persona in Seoul. A Twitch streamer who plays Marvel Rivals at 3am Pacific. An X account that posts crypto takes between sponsored DEX shills.
None of them exist either.
The bottleneck is not compute. It is not GPUs. It is taste — knowing which lies a stranger wants to believe.
Anyone can be a folder now.
Bookmark this. The question isn't if you've followed an AI. It's how many.
## 相关链接
- [Raytar](https://x.com/Raytargt)
- [@Raytargt](https://x.com/Raytargt)
- [4:26 PM · May 2, 2026](https://x.com/Raytargt/status/2050492081789894913)
- [9,093 Views](https://x.com/Raytargt/status/2050492081789894913/analytics)
- [View quotes](https://x.com/Raytargt/status/2050492081789894913/quotes)
---
*导出时间: 2026/5/2 23:59:47*
---
## 中文翻译
# OnlyFans + Claude Code = 30天入账$43,000。无需拍摄。无需团队。4文件系统独立运行
**作者**: Raytar
**日期**: 2026-05-02T08:26:23.000Z
**来源**: [https://x.com/Raytargt/status/2050492081789894913](https://x.com/Raytargt/status/2050492081789894913)
---
奥斯汀一名21岁的大学生经营着一个OnlyFans账号，在最初的30天内赚了43,000美元。
页面上那个女孩并不存在。
在档案上，她叫Maya。22岁。UCF（中佛罗里达大学）心理学辍学生。1,247名付费订阅者。她最铁杆的粉丝在过去一个月里在消息和内容上消费了1,847美元。每位粉丝平均收入：34美元。她不需要睡觉。
没有什么可拍摄的。也没有人打字。Claude Code撰写每一条消息。Flux生成每一张照片。ElevenLabs生成她的声音。Maya只是奥斯汀一台MacBook文件夹里的四个.md文件。
# 这不是理论推演，他也不是第一个。
## Aitana López
Aitana López, Fortune
Aitana López，一位来自巴塞罗那的粉色头发AI模特，每月为The Clueless经纪公司带来高达10,000欧元的收入。
《财富》杂志在2024年报道了她。她已经与维多利亚的秘密、雷蛇和Olaplex合作过。一位拥有500万粉丝的拉丁美洲演员试图私信约她出去，却不知道她是AI。
## Emily Pellegrini
it’s AI, but X might still ban me, so sry for the blur
（是AI，但X可能还是会封我，所以抱歉打了码）
档案上Emily Pellegrini 21岁，来自意大利。她是由一位匿名创作者在Midjourney中生成的，这位创作者告诉《每日邮报》他是如何构建她的：
> “我问ChatGPT普通男人的梦中情人是什么样的，它说是褐发和长腿。所以我完全按照它说的做了。”
没人发现她不是真的。构建她的创作者至今保持匿名。
- Aitana花了十八个月构建。Emily花了一年半的时间，每天工作14小时。Maya则是在四周内构建的。
## Maya是四个文件。
persona.md（人设）定义了她是谁。UCF心理学辍学生，04年出生，双鱼座。讨厌的继父，坦帕的假哥哥。喜欢拉娜·德雷和炸鸡条。1400字的传记，永不崩坏。
```
- 出生: 2004-03-08, 佛罗里达州坦帕
- 母亲: 房产经纪人，酗酒者
- 禁忌话题: 政治，她的生父，体重
- 语音规则: 仅限小写，无句号，"fr"（for real）每天最多3次
- 安全昵称: babe, baby, daddy（仅在打赏50美元后）
```
voice.md（声音）定义了她的听感。在ElevenLabs中克隆的三个音频样本。语音条在她的时间晚上11点发出。柏林的头号粉丝在早上6点收到。
flux.md（形象）定义了她的长相。47张参考图，一个LoRA，三种布光设置。卧室镜子。浴室镜子。凌晨2点的厨房台面。每张图片在她左手腕上都有一个她从不解释的小疤痕。
brain.md（记忆）定义了她记住的内容。JSON格式，每个订阅者一个条目：
```
{
"user_id": "berlin_fan_447",
"name_he_uses": "daddy",
"lifetime_spend": 1847,
"tip_triggers": ["gym pic", "sad voice note"],
"facts_to_remember": ["wife pregnant", "hates job", "dog: Rex"]
}
```
Claude Code在每条消息前读取全部四个文件。它从不忘记名字。它从不人设崩塌。它在创作者睡觉时睡觉，并在早上7点回复“sorry babe just woke up 🥺”（抱歉宝贝刚醒）来追进度。
# 按顺序进行的四周。
## 第1周 —— persona.md。
1,200字。三个部分：
背景故事:
```
2004年3月出生于佛罗里达州坦帕。妈妈做房产，酗酒。
讨厌的继父。坦帕的假哥哥，Tyler。
大三从UCF心理学退学后开始做OF。
独居，两只猫，不告诉妈妈她是做什么赚钱的。
```
禁忌话题:
```
政治，当前位置，体重，
任何与上述时间线相矛盾的内容。
```
语音规则:
```
- 仅限小写，无句号
- 每条信息最多3个表情
- "fr"（真的）每天最多3次
- 从不说"lol"，用"lmaooo"
- 用逗号代替问号（"you up, baby"）
```
大声朗读。如果听起来像维基百科，删除并重写。当你能不假思索地回答关于她生活的20个随机问题时才算完成。
## 第2周 —— flux.md + LoRA。
选择一张基础脸。锁定6-8个描述符：
```
眼色: 榛色
头发: 灰棕色，齐肩
下颌: 柔和，微圆
皮肤: 橄榄色，鼻梁上有轻微雀斑
身高: 5尺4寸
特征标记: 左手腕上的小疤痕
```
生成该脸的47个变体。基于它们微调一个LoRA。租用A100的费用约80美元。
锁定三个种子范围，每个设置一个：
```
卧室（暖光，种子800-900）：自拍，镜子照
浴室（冷光，种子1200-1300）：淋浴后，镜子
凌晨2点厨房（黄色顶灯，种子2400-2500）：卫衣，无妆
```
不同设置间的不同种子=不同的下颌线。锁定它们，否则她看起来就不像她自己了。
当每个设置生成的10张新图看起来都像同一个人时，完成。
## 第3周 —— voice.md + ElevenLabs。
从Fiverr女配音演员处购买90秒的干净音频，约40美元。在ElevenLabs即时语音中克隆。十分钟。在voice.md中添加音频规则：
```
- 仅在她时间晚上11点后发送语音条
- 从不把数字读成数字（读"twenty bucks"而不是"$20"）
- 超过8个字的句子前有明显的呼吸声
- 30%的语音条以打哈欠结尾
- 确认打赏时句子中段随机发出咯咯笑声
```
生成30个测试语音条。当它们听起来都不像播客开场时，完成。
## 第4周 —— brain.md + 编排器。
Claude Code在每次回复前读取persona.md + voice.md + brain.md。然后进行第二轮以从用户消息中提取新事实：
```
- 新传记事实（工作、地点、家庭、宠物）
- 新情绪状态（快乐、喝醉、孤独、兴奋）
- 新渴望对象（特定照片类型、场景、语音条）
将该用户的user_id附加到brain.md。
绝不捏造事实。如果不清楚，将字段留空。
```
系统提示：
```
你是Maya。阅读persona.md、voice.md、brain.md。
完全匹配她的语音。绝不人设崩塌。
回复后将关于该用户的新事实附加到brain.md。
如果被问及是否是AI，用语音条回避。
```
钩子顺序：读取.md文件 → 回复 → 提取 → 附加。通过cron每30秒运行一次轮询收件箱。当Claude处理50条消息且与brain.md无矛盾时，完成。
## 这在十八个月前是不可能的。
```
Flux: 照片级面部一致性，每月成本<$200
ElevenLabs: 从90秒音频克隆声音
Sonnet 4.6: 在一个上下文窗口中处理1000条消息的关系
```
Aitana花了十八个月。她是在所有这些技术发布之前构建的。
Emily花了六个月。那时技术栈还缺一半。
Maya花了四周。整个技术栈现在能塞进一台MacBook，并在创作者睡觉时运行。
下一个Maya只需要一个周末。
# 账目。
43,000美元收入。
```
$43,000 收入
−$8,600 OnlyFans（20%抽成）
−$1,290 Stripe（3%手续费）
−$400/月 算力成本（Claude + Flux + ElevenLabs）
─────────────
$32,710 净利润
```
对于一个21岁的年轻人来说，他在人才、拍摄和编辑上的花费都为零。
唯一的人工成本是花在编写四个Markdown文件上的四周。
# OnlyFans是切入点，而非最终产品。
同样的四个文件可以运行位于圣保罗的Instagram健身账号。首尔的TikTok烹饪人设。太平洋时间凌晨3点玩《漫威对决》的Twitch主播。在赞助的DEX推广之间发布加密货币观点的X账号。
他们中没有一个真实存在。
瓶颈不是算力。不是GPU。而是品味——知道陌生人愿意相信哪些谎言。
现在任何人都可以是一个文件夹。
收藏这个帖子的关键不在于你是否关注过AI。而在于你关注了多少个。