Peak claim: “The only UI generation platform which understands your codebase.”
Final outcome: Shut down in July 2025 after burning $500,000, with the last pivot stuck at $1,500 MRR.
Lesson: If strangers will not pay, another pivot will not save you.
💰 The Money Trail
They did not raise on Demo Day. Vedant Agarwala, cofounder and CTO, said they burned through the $500,000 they had raised. No priced round, no valuation, and no named lead beyond the YC batch showed up in the shutdown coverage. Primary YC partner on the company page: Brad Flora.
👥 The Founder Story
Two builders. A team of two, per YC.
Royal Jain, cofounder and CEO. IIT Bombay computer science. Founding engineer at Deepaffects, later acquired by RingCentral. Before CodeParrot he was CTO of a startup the YC bio describes as $7 million ARR, acquired by an edtech unicorn. The bio does not name that company. That $7 million is not CodeParrot revenue.
Vedant Agarwala, cofounder and CTO. Led backend and mobile teams at Apna and, on the YC bio, scaled systems to 100 million daily requests. He had already dropped out once to start Instano, a quick-commerce company acqui-hired by MagicTiger.
They got into YC Winter 2023 with the founding team and an idea they scrapped within weeks. Vedant’s account: they left India, sorted US visas in Singapore, and spent three months in San Francisco. Then Demo Day came and went with no round.
His shutdown note, 16 July 2025:
“We didn’t raise on Demo Day. We spent the next year stuck in pivot hell, searching for what our company really wanted to be. We hired and, painfully, had to let go of both the engineers. We burned through the $500k we’d raised — and when we hit $1,500 MRR with our final pivot (Figma-to-code using LLMs), we couldn’t break through. Instead of running more experiments with dwindling runway, we decided it was time to shut shop.”
He also wrote: “Good prompts get you 90% there, but good evals are what really matter. And even when you’re building with cutting-edge AI, 70-80% of the codebase is still good old-fashioned software — and it deserves just as much attention.”
🔧 What They Actually Built
CodeParrot was a developer tool that changed jobs before it had a customer who would pay for one of them.
First version: API testing from production traffic. The pitch was to capture live API calls, including downstream calls to databases and third-party services, and cut the boring test-writing. The platform sat in beta for about a year, with talk of an open-source project in 2024.
Next version, and the one they died on: design to code. A VS Code plugin. Point it at a Figma component or a screenshot, set your code preferences once, get production-looking UI. Support reported for React, Flutter, and HTML. The December 2024 homepage told full-stack developers, frontend engineers, and founders they could build new screens from components they already had.
Pro was $19 per seat. Team was $39 per seat, minimum three seats. Enterprise was “get in touch.”
The homepage called CodeParrot the only UI tool that understood your codebase. The founder’s number for the peak of that final pivot was $1,500 MRR.
Two and a half years. One batch. A plugin.
📉 The Wind-Down
This was a choice, made while there was still a little runway.
16 July 2025: Vedant posts that they are shutting down.
17 July 2025: Inc42 reports the shutdown.
19 July 2025: Entrepreneur India reports that CodeParrot has ceased operations.
YC company directory: Status Inactive.
6 October 2026: https://codeparrot.ai does not serve a product.
No reporting found of a buyer, a bankruptcy docket, or money sent back to investors. Investor outcome: unknown. Vedant said he was taking a short sabbatical and would help other YC companies get LLM apps into production, while looking for a cofounder for the next company. Royal’s next step: unknown.
📊 The Obituary Scorecard
Dimension | Score | Reasoning |
|---|---|---|
Product-Market Fit | 1/5 | $1,500 MRR after two and a half years means almost nobody paid. |
USP | 2/5 | “Figma in, code out, in your stack” was a clear sentence, and it was a feature. |
Timing | 2/5 | They rode the AI coding wave and the wave turned their product into a checkbox inside bigger tools. |
Founder Fit | 3/5 | Both had already shipped real systems. They could build. They could not stay on one customer. |
Team (Execution) | 2/5 | They shipped, hired two engineers, then cut both. The year went to pivots. |
💡 Lessons for Solo Founders
$1,500 MRR is a number you can hold in your head. After two years it is a prototype with a Stripe account. It is not a company. Stop, or change the customer. Do not change the slogan.
Hire after a stranger pays, and pays again. They hired two engineers and had to let both go. A solo founder who skips that hire keeps the runway CodeParrot spent searching for what the company was.
Pivot hell is a calendar, not a personality. One job, one price, ninety days. If the same people will not pay, the next idea needs a new conversation, not a new repo on the old one.
A plugin that turns a design into code is a feature a model lab or an editor can ship on a Tuesday. Your wedge has to be a workflow they will not bother to own. “Understands your codebase” was the right sentence. It still has to clear a price.
YC gets you into the room. It does not get you a round. They did not raise on Demo Day. Plan the company as if the room is polite and the wire never comes.
Shut down on purpose. They did. They said the number out loud. Copy that part. A clean stop beats a sixth pivot on fumes.
Vedant’s technical lesson is the one builders skip: prompts get you most of the way, evals decide if the product is real, and most of the code is still ordinary software. A solo founder can ship the wrapper in a weekend. The evals and the billing are the business.
🪦 Final Tally
Founded: 2022, Bengaluru. YC Winter 2023.
Peak MRR: $1,500, on the final Figma-to-code pivot.
Total raised: $500,000. Demo Day round: none.
Team: two founders. Two engineers hired, then let go. Peak headcount: unknown.
Price on the archived site: $19/seat Pro, $39/seat Team.
Shutdown announced: 16 July 2025. News: 17 and 19 July 2025.
Founder outcome: Vedant on a short sabbatical, helping YC teams ship LLM apps, looking for a next cofounder.
They wanted to be the copilot that turned a Figma file into a pull request. The market paid $1,500 a month for the attempt. Then they stopped. That last decision is the grown-up one. The expensive one was the year spent searching for what the company wanted to be, while the price on the homepage never showed up in the account.
A company that already charges
CodeParrot’s cause of death was no product-market fit. The last idea was a generic coding feature, and it capped at $1,500 MRR.
TrustMRR is where a solo founder can look at startups that already have that fit. Revenue there is verified through Stripe, RevenueCat, Superwall, the App Store, or Creem. It is not a founder screenshot.
One listing that does the opposite of this death: Adly. The listing describes a marketing workspace for Arabic-speaking businesses. One kind of buyer. One job. Stripe, synced 5 October 2026, shows about $4,128 in the last 30 days, $3,775 MRR, 75% profit margin, 54 active subscriptions, revenue up about 25% and MRR up about 8% over 30 days, at a 1.9x revenue multiple.
That is already more than twice CodeParrot’s best month, on a niche a single builder can understand. Study the wedge before you write another plugin.
Browse the rest of the board: TrustMRR.
I hope this story serves as a reminder that even a sharp product dies when nobody keeps paying for it, and that staying on one customer matters as much as shipping the next pivot.
If you found it helpful, pass it on. Thank you.
Cheers,
Ram
👉 My simple ask: It took hours to put together this post for you. I hope you forward this email to at least one founder friend or share on your social channels 🙏.
Startup Obituary is for educational purpose only not a business advice.