Indie Hacker Revenue: A Graded Evidence Table

An indie hacker should choose ideas by graded revenue evidence, not forum confidence. ProvenStartups matched 213 solo-run projects; among the full matching set, 65 publish a clean monthly figure, with a $15K/mo median and a $6/mo to $300K/mo range. Start with the smallest product tied to proven distribution, and reject impressive numbers when the source class cannot survive inspection.

Contents

Use this page in order: read the graded revenue table, understand the contradiction, select a narrow model, build a paid test, then verify every claim before copying it. An attractive market story is useless when its revenue source, costs, period, or operator scope is unclear.

• ·The indie hacker revenue table
• ·Where the data contradicts the popular claim
• ·What we would build
• ·How to run the first paid test
• ·How to verify an indie hacker claim
• ·FAQ

The indie hacker revenue table

The usable answer is a revenue table where every claim carries its source quality. [V] means third-party verified, [F] founder-reported, [C] creator-relayed, and [U] unverified. Those labels do not declare a business good or bad; they tell you how much confidence to place in its published number.

This is not a collection of forum self-reports presented as equivalent facts. The amount, evidence class, category, and build difficulty stay visible together, so a large relayed estimate cannot quietly outrank a smaller verified result.

These are comparison cases, not the inputs shown in place of the full cohort. The cohort median uses all 65 clean monthly figures. Its disclosed evidence split records 25 [V], zero [F], zero [C], and zero [U]; because that accounts for only part of the 213 projects, the remaining cases should not be silently treated as verified.

Where the data contradicts the popular claim

The popular claim says indie hackers win by building another small SaaS and letting code quality compound. The data contradicts it: revenue appears across apps, services, publishing, tools, and plugins, while technical difficulty does not track evidence quality. Distribution and proof are the constraints; shipping code is only one input.

In the full cohort, Consumer App and SaaS each contain 32 projects, AI Service has 31, Digital Publishing has 27, and 20 are cautionary tales. Site-wide, ProvenStartups files 38 documented failures alongside the wins. This is not a SaaS-only leaderboard or a survivor showcase.

The most impressive claim is not automatically the best benchmark. nano-banana.ai reports ≈$115K/mo net profit for one month [C], while Peptide Tracker App reports $11K MRR [V]. The first is larger; the second has stronger evidence. Collapsing both into “successful AI apps” destroys the distinction that matters.

Coding tools also appear frequently, but frequency is not causation. Across the whole index, 211 distinct projects mention at least one AI coding tool; ChatGPT appears in 100, Claude Code in 50, Cursor in 46, and Bolt in 40. Stack Overflow's annual developer survey provides broader developer context, not validation of these businesses.

What we would build

We would start with a narrow painkiller sold through an existing channel, preferably a plugin, ecosystem tool, or productized service. We would refuse a broad consumer build, an ad-funded traffic bet, or a mass-produced app portfolio until one channel produces a paid signal whose source and economics can be recorded.

• ·Start beside an existing workflow. Data Fetcher's $23K/mo and 85% margin [F] make a platform plugin worth testing, but the founder-reported grade still requires source review.
• ·Use service work to expose demand. The AEO Service reached a $2,000/mo retainer from one client [F]. One client is validation of a paid problem, not proof of a scalable market.
• ·Prefer observable economics. HabitKit reports $15K MRR against $200–$300/mo in costs [F]. Those disclosed costs make the claim more useful than revenue alone.
• ·Reject modeled upside as operating truth. WordUnscrambler's estimated $170K–$660K/mo [C] comes from traffic and RPM assumptions. It is a research lead, not a revenue statement.

The filter is simple: a reachable buyer, a recurring job, a distribution surface, and a claim you can type precisely. If the plan needs viral luck before the first payment, we would not build it.

How to run the first paid test

Build the evidence trail at the same time as the product. A viable first cycle has one buyer, one job, one acquisition path, and one auditable payment event. The goal is not an impressive launch screenshot; it is enough traceability to separate repeatable revenue from a temporary spike or estimate.

1. 1.Write the claim first. Name the buyer, painful job, price, delivery period, and acquisition channel in one paragraph.
2. 2.Sell the smallest complete outcome. Ask for a paid pilot or prepayment before adding dashboards, agents, or integrations.
3. 3.Ship only the paid path. Instrument lead source, activation, payment, refund, variable cost, and support time.
4. 4.Grade the result. Preserve receipts or third-party records as [V]; label your own disclosure [F]; label retellings [C]; use [U] when the source cannot support the number.

Selling Shovels in the OpenClaw Ecosystem reached $40K in subscriptions in two weeks [C]. That is evidence of fast demand, not automatically stable MRR. Social Wizard + Clean Eats reached $1.5M across two apps in 12 months [F], but it still does not prove that a clone inherits the same distribution.

How to verify an indie hacker claim

Treat every revenue claim as a typed value, not a string. Store amount, period, gross-versus-net basis, operator count, source owner, and evidence grade separately. If a field is missing, preserve the gap instead of coercing the claim into MRR; that discipline is what makes comparisons useful.

The full project index contains 406 graded ideas: 57 [V], 184 [F], 121 [C], and 44 [U]. Of those, 246 are solo-run and 266 are software or SaaS. Read the grading method before treating a headline as a benchmark.

Use third-party context only for the question it actually answers. The Census Bureau's nonemployer statistics can frame businesses without employees; it cannot verify a specific app's sales. ProvenStartups gives Peptide Tracker App's $11K MRR [V] more evidentiary weight than a larger unsupported claim.

FAQ

The short answers below cover the next decisions: definition, expected revenue, idea selection, and coding-tool choice. None substitutes for reading the underlying source. A compact number can guide research, but its evidence class determines how aggressively it should influence build time, positioning, or cash.

What is an indie hacker?

An indie hacker is an independent builder who creates and operates an internet business without relying on a conventional venture-backed team. For this analysis, the operational definition is narrower: all 213 matched projects are solo-run. The label describes how the product is operated, not whether its revenue is verified or repeatable.

How much do indie hackers make?

There is no defensible universal income figure here. Among the full matching cohort, 65 projects publish a clean monthly number; the median is $15K/mo and the range is $6/mo to $300K/mo. Treat those as cohort statistics, then inspect individual grades, such as Peptide Tracker App at $11K MRR [V].

Which indie hacker idea should a developer start?

Start with the smallest paid problem attached to a channel you can already reach. A plugin resembles Data Fetcher at $23K/mo [F]; a service resembles the AEO offer at $2,000/mo for one client [F]. Copy the constraint and acquisition surface, not the finished product or its headline revenue.

Does the AI coding tool determine success?

No. Tool mentions show what builders used, not what caused revenue. Across the full index, ChatGPT appears in 100 projects, Claude Code in 50, Cursor in 46, and Bolt in 40. Pick the tool that shortens the paid test, then evaluate distribution, retention, costs, and evidence independently.