Checklist: When an AI Side Project Becomes a Company

You built it on weekends, with a day job still on the calendar. Now the repo has users, a few invoices exist, and a friend asks if you should incorporate. The honest answer is usually: wait until the project can fund what comes next. Fish Audio closed a $52 million seed round led by Coreline Ventures and Capital Today.

Most side projects stay side projects, and that is fine. The useful question is when to quit your day job, and the answer is when paid demand repeats and revenue can fund the next build before outside money arrives.

Incorporate only when the project passes the checklist

Use the list below as a test. Pass the checks and the project is asking to become a company. Fail them and keep it as a side project, protecting your evenings.

• Open-source traction: your repo shows steady external interest, not just your own commits.
• Paid demand: a handful of users pay for the same job, not a favor.
• Revenue before capital: your income can cover the core costs of the next build.
• Paid enterprise tier: a higher tier promises measurable reliability, not just extra capacity.
• Team and sales readiness: you can answer sales questions without hiring a full team.

Traction and paid demand are the early tests

Open-source attention matters because it shows people outside your circle are willing to spend time. Fish Audio's Fish Speech open-source project had more than 31,000 GitHub stars and pulled in indie developers, video game designers, and creators.

Attention alone does not pay. A paid user signals more than a star, because the user has already spent money on the problem.

Repeatable paid demand is the stronger signal. In its first year, Fish Audio had $21 million in annual recurring revenue and over 8 million users.

Series A investors expect repeatable demand before expansion funding.

Users who ask for invoices, not just for help, show the demand. A repeated workflow in support messages indicates the project is doing a job. Interest only from friends points to hobby mode.

Revenue before capital is the real control

Outside money can speed a build, but it can also create a scale you have not earned.

Revenue before capital gives you the right to say no to a bad round. It also proves your pricing is real, because someone paid for it. With only free users, the next step is a pricing test, not a company formation.

Venture capital reaches only about 1% of startups, and chasing it too early can pull founders away from product-market fit, customer value, and profitability.

UNest's founder used $160,000 of family savings to build the first version before seeking outside capital. UNest then raised a $2 million venture round after the bootstrapped phase.

Customer-funded growth can come from service contracts, subscriptions, paid pilots, retainers, deposits, pre-orders, or payments. A bootstrapped raise can draw on customer revenue, disciplined costs, advance sales, modest checks, grants, revenue-linked capital, and later investor rounds.

Bootstrapped puts customer money before outside money.

Startups on Carta's platform closed 1,187 new venture rounds in Q2 2025, a 13% year-over-year decline.

Fish Audio's CEO said outside capital was unnecessary while the startup was an open-source project for creators. The company sought funding to build more advanced models and serve enterprises as investor interest grew.

Enterprise tier and sales readiness decide expansion

Enterprise buyers need promises they can test. Fish Audio's paid plans include guarantees for latency and uptime. HeyGen and LiveKit were among companies running on Fish Audio's APIs.

Enterprise readiness shows up in answers about reliability and sales.

When the project starts serving larger customers, sales becomes a real workload. Fish Audio said it would use the seed to build an enterprise sales team.

Sales readiness looks like taking a call, identifying the buyer, and following up without adding more unpaid work.

Answer those readiness questions, explain the reliability tier, and keep the revenue steady, and the side project has earned the transition. Otherwise, keep the project where it is and revisit the checklist when paid demand changes.