Ask what a category complains about, what it wants, and what it already pays for. One query goes out across forums, communities, video and news, and the threads come back as structured records with their comment trees attached.
Ask what a market pays for
A market research API returns the public posts, forum threads and comments where buyers describe what they need, what they already use and what they pay for. SocialCrawl runs one query across 46 platforms and returns the matching threads with their comment trees on one schema, so a demand study becomes a script you run rather than fieldwork you commission.
Every one of these runs in production today. The endpoint chain under each card is the actual call sequence.
Phrase the query the way a frustrated buyer would phrase it, not as a keyword, and get the threads where people describe the problem in their own words. The queries that work in production read like sentences: too expensive, unreliable, hard to set up.
search/forums (what people complain about)Pull the recurring questions around a topic from communities and video in one call. The same output doubles as a content brief and a feature backlog, because both are built from the questions nobody has answered well yet.
prism/audience-questions -> content briefSearch with boolean operators, then pull the whole comment tree on anything that matters. Price objections, workarounds and named alternatives sit in the nested replies, not in the post.
reddit/search -> reddit/post/commentsRun one keyword per call across a competitor set and collect what people say about each. This is the shape most category studies take once the brief is a landscape rather than a single product.
prism/demand-signals (category read)Take the products already selling into the category and collect the public verdict on each. Validation work usually ends in a comparison table, so start by filling one with evidence instead of feature checkboxes.
prism/reputation (competitor landscape)Track how a category is described over months rather than taking one snapshot. A shift in the words buyers use is the earliest readable sign that expectations have changed.
content_analysis/phrase-trends (over time)Four calls take you from a vague hypothesis to something you can quote in a deck. Each step narrows the field.
Reads are priced in credits. Empty result sets are not billed on list endpoints, and failed upstream calls are refunded automatically.
GET /v1/search/forumsSend the question in plain language and get forum and community threads where people describe the problem. This is the cheapest way to find out whether the pain is real.
per callGET /v1/prism/audience-questionsPass a topic and the sources you want, and get the questions a market asks repeatedly. Scope it to communities and video for consumer categories.
per callGET /v1/reddit/post/commentsTake the threads worth reading and fetch every comment and reply. The evidence a stakeholder will accept is almost always a quote from this step.
per pageGET /v1/prism/demand-signalsRun the same read across the competitor set to turn one product question into a landscape view.
per callEach of these is a lane in the same research pass. Follow a link to see that platform's full endpoint surface.
Most teams either commission a panel study or read communities by hand. Here is what changes.
| Capability | SocialCrawl | Typical setup |
|---|---|---|
| Evidence source | Unprompted public posts, written before anyone asked a research question. | Panel answers shaped by the survey that recruited them. |
| Coverage | One key spans 46 platforms, including forums, communities, video and news. | One panel or one community, priced separately. |
| Depth | Full comment trees, where the price objections and named alternatives live. | Top-level posts, summarised. |
| Turnaround | A first category read runs today, as a script. | Fieldwork scheduled in weeks. |
| Pricing model | Never-expire credit packs, priced per call, so a one-off study stays a one-off cost. | Project fees or an annual seat. |
| Traceability | Every finding keeps its source post and URL, so a stakeholder can open the original. | Aggregate charts with the underlying responses locked away. |
A market research API returns the public posts, threads and comments where a market discusses a problem, a product or a price, as structured data you can store, score and quote. It replaces manual community reading and, for many questions, replaces the survey that would have been commissioned instead.
Yes, and it is one of the most common patterns on the API. The workflow is to ask what people currently do about the problem, what they pay for today, and what they complain about in the tool they already have. Three queries and their comment trees usually settle whether the idea has a buyer.
SocialCrawl covers 46 platforms and 375 endpoints, including Reddit, Hacker News, YouTube, X, Trustpilot, news and the open web. One API key covers all of them, and one query can fan out across them together.
A survey collects answers to your question. This collects what people said when nobody was asking, which removes the framing effect but adds the work of finding the relevant threads. In practice teams use it first to find the real questions, then survey only what remains genuinely uncertain.
Yes. Phrase and category trend endpoints show how the language around a category moves month to month, and any search can be re-run on your own schedule and stored. Repeat lookups of the same topic are the most common production pattern on the API.
Reads are priced in credits from never-expire packs, so you pay per call rather than per seat or per project. A typical first pass is a handful of searches and the comment trees underneath them. Signup includes 100 free credits and no card is required.
Yes, for transformed views and analysis you present to your own client or users. Our terms permit derived research products and restrict raw passthrough or bulk redistribution. If you are unsure where your deliverable sits, ask us before you ship it.
Ask AI about SocialCrawl
100 free credits on signup. No card, no sales call, no minimum project size.
curl -G "https://socialcrawl.dev/v1/search/forums" \
-H "x-api-key: $SOCIALCRAWL_API_KEY" \
--data-urlencode "query=project management tool too expensive"curl -G "https://socialcrawl.dev/v1/prism/audience-questions" \
-H "x-api-key: $SOCIALCRAWL_API_KEY" \
--data-urlencode "topic=ai note taking apps" \
--data-urlencode "platforms=reddit,youtube"Both calls run against live data. Swap the key and the query and they work as written.