Social research workspace
Type a query. Read matching records. Save them. Watch the query. Results come from live public posts across 64 platforms you check — native APIs/RSS where they exist, public index for locked networks.
Scrapely searches 64 public platforms you check — social, video, forums, news, reviews, and developer sites. Open sources (Reddit RSS, Hacker News, Wikipedia, GitHub, GitLab, Stack Overflow, Mastodon tags, and others) are fetched natively. YouTube Data API and X (lobstr.io) use keys each signed-in user pastes in Settings — never another account’s keys. Locked networks (LinkedIn, Meta, TikTok, Bluesky without auth) go through the public web/news index, and through SocialCrawl universal search when the operator sets SOCIALCRAWL_API_KEY — still not a paid firehose. Every result keeps a source URL. Empty means no public hit, not a complete firehose.
Author, platform, time, engagement, entities, and a source URL. Save it, add it to an investigation, or run Analyze on the stored records — with citations back to those records, not a chat with no memory.
Author + handle
Who posted, on which network.
Record body
The public text, truncated to what the source returned.
Source URL
Open the original. Provenance is the point.
01
A box on the first screen. Guests get a short live result list from the platforms they check.
02
Account pages, entities, timelines, a relationship graph. Weak links are labeled as hypotheses, not identity.
03
Queries, posts, notes stay on the workspace.
04
Re-run a query. Events appear in the app. Outbound webhooks are not sending yet.
05
Optional, metered, cites stored posts. If the model is down, you get an extractive brief.
06
Unsubscribe is in the footer. Cancel is on Billing. No maze.
Brand mentions and campaign leftovers — on records you are allowed to see.
Complaints and switch-off language in public posts.
Shared monitors and named reports. Not a custom white-label domain.
Public buying language, not a purchased contact list.
Public professional chatter — not an FCRA background check.
Timelines and source URLs. Demo records are not publishable facts.
Operators and graphs with stated confidence. No private-account access.
Seats and shared saved work. SSO is not in this version.
A saved query, re-run. Free has one. Events stay in the workspace. We do not claim outbound webhook delivery yet.
A named pile of searches, posts, and notes with a stable URL in your workspace.
Summaries that must point at stored posts. Credits are metered because model calls cost money.
Paid plans are billed through Stripe. Checkout collects a card; workspace limits update after payment succeeds. Cancel anytime on Billing — keep the period, or drop to Free immediately.
Learn the workflow. Real search, tight limits.
$0/mo listed
The working professional plan. History, AI, reports.
$49/mo listed
Shared monitors, roles, and a workspace that compounds.
$249/mo listed
Platform, source id, source URL, provider, retrieval time. AI must point back.
No CAPTCHA breaking, no private accounts, no unofficial scraping as a feature.
Each search hits public surfaces for the 64 platforms you checked. Native APIs/RSS where they exist. Results are labeled with provider and source URL.