Indie hacking timeline

Indie hacking, on this site, is people making money from products they own. A timeline is how that practice got its public shape. A “state of” only works if you can see which advantage just became a template, and which template the crowd is already copying.

The ranges overlap. Nomad-era products were online before anyone used the name. Build-in-public threads ran while a following turned into a distribution plan. ChatGPT landed in 2022, and starter kits were still the thing people bought in 2023. The walk below uses the same era names and date ranges as the homepage. Dates and revenue are linked. A founder-reported number is one person’s books, posted because posting it was part of the work.

Before the name, 2006–2013

The practice is older than the phrase. On 1 March 2006, 37signals released Getting Real, a short argument for small teams, shipping before the company looked finished, and charging. The book did not use this name. It is the bootstrap-web ancestor the homepage points at when it says the roots were older.

Late in 2013, Ryan Hoover started Product Hunt as a place to share new products. Launches got a public square. That square later became a ritual for small makers. Stripe made taking a card on the internet ordinary for a product this small. The useful change is the one the homepage already uses: payments got easier, so a person with something to sell had a way to charge.

Nomads, 2014–2019

In 2014 Pieter Levels, a Dutch maker, set out to launch 12 startups in 12 months. Most of them did not last. One began as a spreadsheet of good cities for remote workers in Chiang Mai. It became Nomad List. Remote OK followed as a jobs board.

On 30 May 2019 he wrote that the two sites together were at about $83,333 a month, a million dollars a year, without ads or funding. The figure is his. What traveled was the shape: one person, one laptop, recurring internet revenue, no office.

He kept repeating the pattern after those years. On 3 July 2023 he posted that Photo AI, about 14,000 lines of PHP, had 1,872 paying customers and $61,808 that month. Same lesson, later stack. Start with a small useful thing. Charge early. Keep the company light.

A name, 2016

On 11 August 2016 Courtland Allen put Indie Hackers on Hacker News. The title was plain: learn how developers are making money. He had interviewed founders of independent products and published the revenue, not only the launch. The about page still describes the site, created that August, as a place to learn from profitable online businesses and side projects. Many people worked alone. Being solo was never a rule.

Stripe bought the site in April 2017. TechCrunch quoted Allen’s new title as “indie hacker at Stripe,” a small sign the phrase had already become an identity. In April 2023 he and his brother Channing said they were running it again as a standalone business, with Stripe still in as an investor. By then people who never posted on the forum still used the word.

The name arrived in the middle of the nomad years. The laptop picture and the community name share a calendar and are not the same chapter. What indie hacking means follows the definition. Who counts as an indie hacker follows the person.

Build in public, 2020–2023

“Build in public” meant sharing the product, the mistakes, and often the monthly recurring revenue on Twitter, now called X. A private process became a feed. Followers could become early users. The numbers made a story concrete. They were also founder-reported. The audience usually had to take the maker’s word.

In September 2021 Damon Chen posted that he had reached $100,000 in annual recurring revenue. Indie Hackers shared that post on 19 September, and the product the thread circled was Testimonial.to. The share’s headline mentions three projects. Read the milestone as his public claim, not as an audit of one dashboard.

Tony Dinh’s Black Magic is the other exhibit from the same years. In a May 2023 newsletter he wrote that he sold the Twitter tool for $128,000 after new API pricing made it unprofitable. At the sale he put it around $14,000 a month. The habit did not stop with that product. In October 2025 he wrote that TypingMind was making about $130,000–$160,000 a month. Both figures are his.

Marc Louvion, known as Marc Lou, was the serial version of the same feed. Product Hunt’s 2023 Maker of the Year note says he launched 14 products that year. The public streak is the launches. The pile under it is this site’s look at $3 million and 36 startups.

Audience, 2021–2024

The next shift treated the crowd as part of the product. Share the build. Launch on Product Hunt. Post the result. Sell the next thing to people who had watched. Lou’s playbook captured the era. A following lowered the cost of a first launch and gave every new idea a starting crowd.

Attention was not the same as demand. People can enjoy a build and never buy the result. A following can lower the cost of a first launch. It is not a purchase order. In his January 2026 year-in-review, after three launches that did not work, Lou wrote the line this era keeps needing: having an audience does not guarantee a successful startup. Quiet businesses that sold through search, email, or a community they already had dropped out of the argument, because they had nothing to screenshot.

The audience also narrowed. buildinpublic.so argues the performance version of the feed — daily MRR, milestone-as-content — is the part that died. LaunchKit is harsher: for the average maker the window closed around 2024, because the people watching were other builders, not the customer who needed the tool. That is why the next product, so often, was a product for makers.

Boilerplates, 2023–2025

Then makers sold the shortcut. ShipFast went up on Product Hunt on 31 August 2023: a Next.js app with login, payments, and email already wired. Buyers could skip setup. Many similar starter kits followed. For a while, the most reliable customer was another person trying to become an indie hacker. Courses, templates, and starter code turned the method into a product.

By March 2026 the engine had moved. In a public month recap Lou posted $78,120 for the month: TrustMRR at $36K, DataFast at $21K, ShipFast at $9.2K, and CodeFast at $8.3K, plus smaller lines. He wrote that AI had killed his coding course and his boilerplate. ShipFast was still on the list. It was no longer the top line. Those are his numbers for one month.

A 2025 total gets mixed up with that month. The same January 2026 letter is titled as $1,032,000 for the year. Inside it, investing had returned $147,000 and X had paid $14,339. The headline is a portfolio year, not a software-only ARR. The longer story keeps that mix in view.

AI shipping, 2022–2026

ChatGPT opened to the public on 30 November 2022. Within a couple of years, people who could not write much code were shipping small products. Copying got cheaper in the same stretch, so a working idea filled up with lookalikes. The tools arrived while boilerplates were still selling. The bottleneck moved later: from “can I build it?” to “can I reach anyone?”

In February 2025 Andrej Karpathy named the loose version of that workflow. In a post on X he called it vibe coding: describe what you want, accept the generated code, and keep moving. He said he no longer read the diffs. He framed it, then, as fine for throwaway weekend projects. The name escaped the post.

On 30 July 2026 Levels wrote that indie hackers might be the first developers to feel AI take execution away as a moat. He put the new price of building at about $20 a month, and he described customers vibe-coding a private copy instead of paying for a thin tool. That number is his. Read it as a claim about defensibility, not as a price list. In September 2026 he posted a cancel-subscription streak and put the savings at roughly $25,000 a month. What he says he kept paying for is the vibe-coded stack story.

Agents are the next stretch of this era, not a separate homepage chapter. Some products are built so other software can be the user. Anthropic announced the Model Context Protocol on 25 November 2024 as one public socket for that. A scheduling product rebuilt so an agent could be the customer is the Postiz story. Ownership and a sale still mark the business. The buyer might be a program acting for a person.

Verification, 2025–2026

Revenue screenshots had become a way to win attention. Some were fake. Marc Lou’s newsletter dates the call-out to 28 October 2025: he saw a post from Levels about fake MRR screenshots, and the next morning he started TrustMRR. Connect a read-only Stripe key. Show revenue the founder cannot edit by hand. He wrote that the first version was up within hours, and that sidebar ads made $20,378 within three days. The newsletter is his account.

TrustMRR became a directory and a marketplace where people can list startups for sale. It is not a revenue-share ad network. In the January 2026 letter he said he hoped an acquisition fee could replace the ads. The March 2026 recap credited a 3% fee for the month TrustMRR led his list. Verification created its own business: proof, then a fee, then ads beside the proof. The screenshot stopped being enough. The attention around the replacement was still something a sponsor could buy.

Bidding, now

Jonathan Wilke, a maker in Germany, launched outbid.lol on 19 August 2026, at 11:08 p.m. The rule is the product: pay more than the person above you and your link sits at number one. No revenue proof. No API key.

A week-one recap, as written up by Steal What Works, reported about 1.36 million visitors, roughly $221,000 in bid revenue, and more than 300 copycats. Those figures are his report, not a census. A public tracker, afford.bid, was listing more than 350 bidding boards by late August 2026. A live tracker moves.

The idea has old relatives. Alex Tew launched The Million Dollar Homepage on 26 August 2005 and sold pixels for $1 each. Marc Köhlbrugge launched highscore.money in May 2016: pay any amount, and that amount is your rank. The 2026 version is the same instinct after shipping got cheap. Attention is the scarce thing, so the leaderboard sells the top spot. One descendant left the browser: Lou auctioned ad slots on his body for a HYROX race.

Where we are now

This is the state in 2026. Software is cheap to make. Attention is expensive. The live plot is paying to be seen, next to verified-revenue pages and vibe-coded clones. On 23 September 2026 Levels said he had passed $10 million a year. The line under that figure, in his own post, is revenue plus investment gains, not the businesses alone. This site’s note on that mix keeps the two piles apart.

The gold-rush loop — a thin product, a launch thread, a screenshot — is what feels dead. The longer argument is whether indie hacking is dead. The bet under the eras did not move. A person can earn money from a product they own. The practice is that bet. The person is whoever is still making it. The next era starts when people get bored of bidding, and someone finds another way to be seen.

FAQ