Product Hunt Data for Market Research: 6 Jobs

How do you use Product Hunt data for market research?

Product Hunt published an official MCP server on November 9, 2026, which means Claude, ChatGPT and Cursor can now query it directly. We wrote up how BigIdeasDB chat calls it live alongside our own datasets. This guide is the method: what to ask, what comes back, and where each answer stops being reliable.

Top 3 ways to research with Product Hunt data, ranked

Ranked by how far the answer gets a founder toward a build, enter or skip decision. Product Hunt data is only the first half of that decision, so the ranking rewards setups that add demand data in the same reply.

What is Product Hunt data?

Product Hunt data is the public record of products launched on Product Hunt: name, tagline, makers, categories, launch date, community votes, reviewer ratings, review highlights, and for many products company and funding details such as Y Combinator batch, total raised and investors. It is a launch log with early sentiment attached, not a market database.

That distinction decides how you use it. A launch log answers “who is building here, and when?” It cannot answer “how many customers does this category have?” Founders who mistake the first for the second end up reading a quiet launch page as proof of an empty market. We show an example of that trap in job 3 below.

You can read it three ways: browse the site, write code against the developer API, or ask an AI assistant that is connected through Product Hunt's MCP server. The third is the one growing fastest, and it is what the rest of this guide assumes.

What can Product Hunt data tell you, and what can it not?

It answers launch, sentiment and funding questions well, and it cannot answer demand, revenue or density questions at all. The table lists nine common research questions and where each one is actually answered. BigIdeasDB comes first in the “otherwise” column because that is where the missing half lives.

Read the table as a rule: Product Hunt is strong on the supply side (who built what) and silent on the demand side (who wants it and pays). The deeper reading is in our guide to competitor analysis with Claude, which covers the same split for any assistant.

What are the 6 research jobs you can run on Product Hunt?

Each job below is one public Product Hunt lookup. Every output was pulled live on November 12, 2026 and shown as returned. The niche throughout is appointment scheduling for service businesses, chosen because it is a real vertical where launches are easy to find and demand is easy to check.

How do you pair every Product Hunt signal with demand data?

Take each Product Hunt finding and ask the question it cannot answer. For the scheduling niche above, the completed picture on November 12, 2026 looked like this:

Here is the finding only the combined view shows. Product Hunt suggests dental scheduling is empty. The Stripe Index shows only 14 dental companies in total, so the niche really is thin. Capterra shows existing patient scheduling tools frustrate buyers with data migration, which is a switching-cost problem rather than a feature gap. And the revenue screen shows small scheduling startups earn a median of $45 a month. Read together: a dental-specific tool is plausible, but the wedge is migration, not another booking page, and the buyer is a practice, not a hobbyist.

That is the workflow our idea validation guide formalizes. For the money side, start with Revenue Intelligence and the Stripe Index, and for the customer voice behind a gap, see voice of customer research. If your question is specifically who already competes with an idea, the sibling guide how to find competitors for a startup idea goes wider than Product Hunt.

What are 8 copy-paste prompts for Product Hunt research?

Paste these into BigIdeasDB chat, or into any assistant that has Product Hunt connected. Replace the brackets. Each one names the lookup, so the assistant does not guess.

1. Alternatives plus pain: “List Product Hunt alternatives to [competitor] with ratings and review counts, then tell me what customers complain about in [competitor].”
2. Niche by use case: “Search Product Hunt for [job, 3 to 8 words]. For each result give the launch year and whether it has any reviews.”
3. Funding check: “Pull the Product Hunt profile of [competitor]: makers, total raised, latest round, investors and the date of that funding snapshot.”
4. YC radar: “Which Y Combinator-backed products launched in the last 30 days in [category]? Show batch, amount raised and votes.”
5. Category map: “Show the top-rated and the newest products in Product Hunt's [category] category side by side, with funding next to each.”
6. Month theme scan: “What topped the Product Hunt leaderboard in [YYYY-MM]? Group the winners by theme and tell me which themes have matching customer complaints.”
7. Density test: “Take the Product Hunt alternatives to [competitor] and compare them to the Stripe Index company count and crowdedness for the same category.”
8. Build or skip: “For [niche], combine Product Hunt launches in the last 365 days, customer complaints, startup revenue and Stripe Index density. Give me one paragraph on whether to build.”

Prompt 8 is the one that matters. A single-source answer is a summary. A four-source answer is a decision, which is why we built the chat to call all of them, and why connecting the BigIdeasDB MCP server to your own assistant pairs well with Product Hunt's.

How do you track competitor product launches over time?

Track launches with a fixed weekly query, a short watchlist and a log, because a launch only becomes meaningful when you can compare it to last month. The process takes about ten minutes a week.

1. Build the watchlist once. Write down 5 to 10 competitor names and 2 to 3 Product Hunt category slugs for your niche. Use the alternatives lookup to find names you missed.
2. Run one query every Monday. Ask for launches in the last 7 days in each category, and profiles for any competitor name that appears. Pace repeated calls: Product Hunt limits public calls to about 60 a minute across all users of the same connector.
3. Log five fields. Launch date, votes, rating and review count, funding with its as-of date, and YC batch if any. A spreadsheet is enough.
4. Add the demand check once a month. For new entrants, ask what customers complain about in the incumbent they target. A launch that fixes a documented complaint is a real threat. A launch that does not is noise.
5. Review the leaderboard quarterly. Pull each month of the quarter and note which themes repeat. A theme that wins three months running is a trend, not a spike.

Example output from this week: a search for launches matching “scheduling” over the last 365 days returned Toki Coordination (September 16, 2026, 217 votes, $11M raised, 5.0 on 17 reviews) and Donna (September 25, 2026, 48 votes, no funding). The other six results matched the word in a tagline, such as “on a schedule” or “scheduled CSV exports”, which shows that launch search is a literal text match. Filter by category slug when you want precision.

If you want the headline launches without running anything, our free LaunchDB weekly launch board covers them. Use the chat for the filtered, repeatable version. For a wider set of tools to run next to it, see best competitor research tools for founders and best Product Hunt alternatives.

How do the Product Hunt MCP and API fit into research?

Both are ways to read the same public data. The MCP is for asking in plain English from an AI assistant. The API is for writing code. For research, choose the MCP.

MCP setup. Product Hunt publishes an install page for Claude, ChatGPT and Cursor. Follow its steps for your assistant. You do not need a Product Hunt account for the public research tools. In BigIdeasDB chat the connection is already built in, so there is nothing to install, paste or authorize. If you want both sources in your own assistant, add the BigIdeasDB MCP server next to it and ask one question that spans launches, complaints and revenue.

API. Product Hunt documents a GraphQL API for developers. It suits dashboards and scheduled exports, but it asks you to handle tokens, pagination and schema. The MCP returns structured results (ratings, makers, funding, launch date) from one sentence, which is faster for a founder doing market research and not building a pipeline.

What are the limits of Product Hunt data?

Seven limits matter, and every one showed up in the pulls above. Treat these as a checklist before you quote any Product Hunt number.

1. Top 8 per call. Every lookup returns at most 8 products. The 7-day and 30-day YC pulls returned identical sets because of this. Narrow by category or query to go deeper.
2. Early-adopter rating skew. Reviewers are makers and early adopters. Eight MINDBODY alternatives all scored 5.0 on 15 reviews combined. A rating on fewer than about 20 reviews is an anecdote.
3. Unrated is not zero. Seven of eight dental results and seven of eight newest scheduling launches carry no reviews. Product Hunt shows 0 and 0, but the right reading is “unrated,” not a verdict of zero quality.
4. Fuzzy name matching. On November 12, 2026, “procore” resolved to Procoto, “clio” to Clip and “housecall pro” to HouseCall. Always check the original-product name in the response, and use the exact product name or slug.
5. Funding is the parent company's, and dated. Reddit Reviews showed over $1.9B because the field describes Reddit. Linear's funding snapshot was dated May 28, 2026. Verify rounds elsewhere before you cite them.
6. Rank is not vote order. Alsona (500 votes) ranked below Wanaka Studio (495). Quote Product Hunt's rank, not a re-sort.
7. Launches are not companies. A category with few launches can still have thousands of operating businesses. Use density data to check.

Last, the rate limit: roughly 60 calls a minute shared across users of a connector. If a lookup is refused, wait a minute. In BigIdeasDB chat the refusal is passed back and the chat still answers from our own data.

Methodology and data sources

Every number in this guide comes from one of the sources below. Product Hunt lookups were run one at a time on November 12, 2026, and results are shown as returned.

Already have a market in mind? Scan it for complaints, revenue and competitors before you write a line of code.

Scan a niche to build in →

Two studies built on the same data: what wins Product Hunt in 2026 and whether Product Hunt is worth it for B2B SaaS.

FAQ