Mara Chen’s second monitor still glows with the dashboard that paid her rent for four years. Not the Stripe metrics for Flowstate—the workflow automation SaaS she founded in 2021 and recently took to $1.2M ARR—but the Chaturbate analytics panel she hasn’t fully let go of. “It’s my calibration tool,” she says, toggling between a cohort retention chart and a token-velocity heatmap from 2019. “Every feature we ship, I ask: would this have saved me three hours on a Tuesday night when I had 2,400 viewers and a broken tip menu?”
Chen, 31, belongs to a quietly swelling cohort: former top-percentile adult creators who have pivoted from performing on the rails to laying them. Their startups—tax-compliance tools, clip-store aggregators, DM automation, multi-platform scheduling—are solving problems mainstream SaaS founders only discovered after OnlyFans’ 2021 policy whiplash made “creator economy infrastructure” a pitch-deck staple. But Chen’s trajectory, from a bedroom in Sacramento to a seed round led by a generalist venture firm that still asks her to redact “adult” from the deck, exposes the fault lines between the creator economy’s rhetoric and its plumbing.
“Investors love the TAM slide for ‘creator tools.’ They just don’t want to know the TAM pays its bills with live interactive webcams and custom clips.”
— Mara Chen, founder, Flowstate
The Economics of Attention: Why the Top 0.1% See Different Problems
Before Flowstate, Chen was “MaraV”—a consistent top-50 earner on Chaturbate and MyFreeCams, grossing $42,000–$58,000 monthly at peak. XBIZ’s 2022 earnings survey placed the platform’s top 100 models at a median $18,400/month; Chen was 3.2x that. Her operation resembled a micro-media company: three moderators on rotating shifts, a clip pipeline feeding ManyVids and OnlyFans, a Snapchat premium tier, and a Discord community she monetized via Patreon before Patreon tightened adult-content enforcement.
“People think camming is performative labor,” Chen says. “It’s operational labor. I was running a 24/7 live-TV station with one camera operator—me—and zero budget for tooling.”
The grind produced a specific dataset: she knew exactly which creator video archives converted subscribers, which tip-menu items had the highest LTV, and how chat sentiment correlated with next-week retention. “I had better cohort analysis on my fans than most Series A B2B SaaS companies have on their enterprise accounts,” she says. “Because if I got it wrong, I didn’t miss a quarter—I missed rent.”
The Pivot: From Feature Requests to Product-Market Fit
The genesis of Flowstate wasn’t a hackathon. It was a Notion page titled “Shit I Pay Humans to Do” that grew to 47 line items: caption writing, thumbnail A/B testing, cross-posting to 11 platforms, DM triage, tax-category tagging, chargeback evidence compilation. Chen hired a dev shop in Buenos Aires to automate the first three. When the bill hit $8,400 for a buggy MVP, she learned React in six weeks and rewrote it herself.
“The insight wasn’t ‘creators need scheduling.’ Buffer exists. The insight was: creators need scheduling that understands a tip-menu dependency graph,” Chen explains. “If I schedule a ‘cumshow’ at 9 p.m., the clip gets cut at 9:22, the thumbnail needs the token-goal overlay, the OnlyFans teaser drops at 9:30 with a swipe-up to the full VOD, and the tax tag is ‘live-show revenue’ not ‘subscription.’ Generic tools treat these as independent tasks. They’re a state machine.”
Flowstate’s core IP is a platform-agnostic event bus that ingests webhooks from Chaturbate, Stripchat, OnlyFans, Fansly, ManyVids, and Patreon, normalizes their disparate event schemas (tip, subscription, purchase, message, ban), and triggers workflows: auto-clipping, caption generation via fine-tuned Llama-3, multi-platform publishing, and QuickBooks-ready ledger entries. The product launched in private beta to 40 creators Chen knew personally. Twelve months later, 38 were paying $199–$499/month. Net revenue retention: 142%.
“We don’t sell time-saving. We sell compliance survival. When a creator gets a 1099-K for $280k and their accountant asks for ‘proof this $12k chargeback was fraud,’ Flowstate produces the audit trail in four clicks. That’s the renew.”
— Mara Chen
The Adult-Adjacent Fundraising Wall
Flowstate’s $2.1M seed round (SAFE, $12M cap) closed in March 2024 after 14 months of “polite ghosting.” Chen’s deck listed Forbes’ $250B creator-economy TAM estimate and AVN’s coverage of adult-adjacent SaaS valuations. The feedback loop was consistent: “Love the metrics. Can’t get IC approval on the ‘adult adjacency’ risk.”
“One partner at a top-10 firm told me, ‘If you rebranded as ‘passion-economy workflow for fitness instructors,’ I’d write a term sheet tomorrow,’” Chen recalls. “I said, ‘Fitness instructors don’t have chargeback rates of 1.8% or need OFAC screening on every payout. My customers do. That’s why the product works.’”
The round eventually came together via a syndicate of angel investors who were former creators themselves—including a former OnlyFans top-0.01% earner who exited a clip-store roll-up—and two micro-funds with explicit “vice-clause carveout” mandates. Crunchbase lists the round as “Seed - Adult Industry Software.”
Platform Policy as Product Strategy
Flowstate’s roadmap is effectively a derivative of platform ToS changes. When OnlyFans briefly banned “sexually explicit conduct” in August 2021, Chen’s beta users lost 37% of revenue in 72 hours. Flowstate’s “Platform Diversification Scorecard”—a feature that calculates a creator’s revenue concentration risk and auto-generates migration playbooks—shipped three weeks later.
“Mainstream SaaS founders read TechCrunch to learn about API changes,” Chen says. “My users are the API change. When Chaturbate killed its affiliate program, my customers’ acquisition funnels broke. We built a referral-link tracker that works across six cam sites in two days. That’s the moat: latency between policy shock and product response.”
This latency advantage compounds. Flowstate now processes ~2.3M events/day across 1,100 paying creators. Its churn is 1.8% monthly—half the median for vertical B2B SaaS per KeyBanc’S 2023 SaaS Survey. LTV:CAC sits at 9.4:1.
The Stack Beneath the Stack
Chen’s architecture choices reveal the constraints of adult-adjacent infrastructure. No AWS GovCloud (too expensive). No Stripe (prohibits “adult content services” per Restricted Businesses list, section 4.3). Flowstate runs on Hetzner dedicated servers in Nuremberg, uses Paddle for merchant-of-record billing, and maintains a custom OFAC/sanctions screening pipeline because Paddle’s built-in compliance flags 12% of creator payouts as “high risk” (mostly false positives on Eastern European names).
“We spend 18% of engineering cycles on payments plumbing that a generic SaaS would outsource,” says CTO Diego Ortega, who joined from a fintech unicorn. “But that plumbing is the product. If a creator’s $40k monthly payout gets held for 21 days, they don’t tweet about our uptime. They churn.”
Data privacy is another differentiator. Flowstate stores zero raw chat logs or media files—only metadata hashes and workflow state. “My users have been doxxed, stalked, had content leaked,” Chen says. “If we get subpoenaed, there’s nothing to hand over but encrypted event IDs. That’s a feature we sell.”
From Insider to Institution: The Next Phase
Flowstate’s Series A deck (target: $8M on $40M pre) adds a new slide: “Platform Risk as a Service.” The thesis: every vertical SaaS serving creators—whether for podcasters, educators, or fitness coaches—will eventually face the policy volatility that adult creators have navigated for a decade. Chen wants to white-label Flowstate’s event bus, compliance engine, and diversification toolkit to horizontal creator platforms.
“Substack, Patreon, Kajabi—they’re all one policy change away from the same chaos we live in,” Chen says. “We’ve been stress-testing this infrastructure for years. The moat isn’t the code. It’s the scar tissue.”
She still camps on Chaturbate once a quarter. “Not for money,” she says. “For calibration. The chat moves faster than any user-research panel. If a feature doesn’t make sense at 2 a.m. with 3,000 people typing, it doesn’t make sense.”
The second monitor flickers. A token goal hits. The heatmap spikes. Chen minimizes the browser tab, opens the Flowstate dashboard, and watches the webhook fire: event_type: tip_goal_reached, amount: 5000, trigger: auto_clip_initiated. The system works. She built it for herself. Turns out, 1,100 other people needed the exact same thing.