Track trends across Reddit, TikTok, and YouTube from one topic query
Last updated
Datapika's trend intel actor runs a single topic query across 13+ platforms, including Reddit, TikTok, Instagram, YouTube, Hacker News, GitHub, Polymarket, Threads, Pinterest, and Bluesky, over a window of 7, 14, 30, or 90 days. Every result is scored by real engagement (upvotes, views, likes) and merged into cross-source clusters, so the same story on Reddit and YouTube becomes one insight. From August 31, 2026 the actor is free to run, with only your Apify platform compute usage billed; you still choose a report depth from Quick Scan to Full Brief, with Standard as the default. Optional AI synthesis adds an executive summary, key findings, sentiment, and momentum, delivered as JSON or Markdown with no API keys required.
How does cross-platform trend analysis work from a single query?
You enter one topic, such as a product name, a person, a company, or a phrase like "best CRM tools 2026", pick a timeframe, and optionally restrict the sources. The actor searches every platform in your tier and returns one dataset instead of ten separate exports. Ranking uses the engagement the platforms themselves report, so a Reddit thread with 2.1K upvotes outranks a thin page that merely ranks well in search.
Cross-source clustering is what saves analyst time. When the same story surfaces on Reddit, Hacker News, and YouTube, the actor merges those hits into one cluster and reports how many platforms carried it. The README's example output for "OpenAI vs Anthropic" shows 47 results across six sources: 15 from Reddit, 10 from Hacker News, 8 from TikTok, 6 from YouTube, 5 from GitHub, and 3 from Polymarket.
- Input: one topic string, a timeframe of 7d, 14d, 30d, or 90d, and an optional list of sources
- Per-platform payloads: Reddit posts with upvotes and subreddit breakdown, TikTok and YouTube videos with view counts and transcripts, GitHub repos with stars and releases
- Dataset row fields: total_results, source_counts per platform, cluster_count, tier, price, and generated_at, with ranked_candidates and clusters in the full report
- Polymarket adds prediction-market odds backed by real money, useful for events and launches
- Every run fetches live data, so a report reflects today's engagement numbers
Which report tier should you use for trend monitoring?
The four tiers differ in which platforms are queried and how much is pulled per platform, not in output format. Quick Scan covers Reddit, Hacker News, Polymarket, and GitHub, which is enough for a daily pulse on technical or financial topics. Standard adds TikTok, Instagram, and YouTube and is the tier most teams should start with.
Deep Intel keeps the same platforms but fetches more results per source plus comment threads and video transcripts, which matters when you want quotes rather than headlines. Full Brief adds AI web grounding so the synthesis can cite pages outside the ten social sources. As of August 2026 the actor has logged 215 runs on the Apify Store. From August 31, 2026 there is no per-report charge for any tier: a run costs only the Apify platform compute it consumes, and deeper tiers use more compute because they fetch more.
- Quick Scan: free-tier sources only, best for high-frequency checks
- Standard: adds the three short-video and image platforms where consumer trends start
- Deep Intel: more results per source plus comments and transcripts for quote mining
- Full Brief: everything above plus web grounding for the AI summary
- No per-report fee from August 31, 2026; the only cost is Apify compute usage, so tier choice is about depth, not price
What does the AI synthesis add to a trend report?
With aiSynthesis set to true (the default), the actor writes an executive summary over the clustered results and extracts key findings, each tagged with its source platform and engagement figure, for example a finding sourced from Reddit with 2.1K upvotes. The README's output example also classifies overall sentiment as bullish, bearish, neutral, or mixed, and momentum as accelerating, steady, declining, or emerging.
The momentum label is the field that makes the actor useful for monitoring rather than one-off research. Run the same topic weekly and the momentum value tells you whether attention is building before the raw counts make it obvious. If the synthesis step fails for any reason, the run still returns the full scored and clustered dataset and, from August 31, 2026, costs nothing beyond the compute the run used.
- executive_summary: a short narrative over the top clusters, written for a reader who has not seen the data
- key_findings: an array of findings, each with a source platform and an engagement number
- sentiment: one of bullish, bearish, neutral, or mixed
- momentum: one of accelerating, steady, declining, or emerging
- best_takes: the highest-engagement quotes across platforms, useful for content and sales prep
How do you track trends across Reddit, TikTok, and YouTube on a schedule?
Schedule the actor on Apify with a fixed topic and a 7d timeframe, and each run becomes one row in a time series. From August 31, 2026 the actor itself is free, so a weekly Standard report or a daily Quick Scan costs only the Apify platform compute each run consumes, which makes high-frequency monitoring on several topics practical. Choose JSON output when the rows feed a dashboard or a database, and Markdown when the report is emailed to people.
For agents, the same actor is exposed through Apify's MCP server, so an assistant can call it by name, read the momentum and sentiment fields, and decide whether to alert a human. The sources filter lets you narrow a scheduled run to the platforms that matter for your audience, for example TikTok and Instagram for a consumer brand, or Hacker News and GitHub for a developer tool.
- Four weekly Standard runs per topic per month, with momentum visible after the second run and no per-report fee from August 31, 2026
- Daily Quick Scan: 30 runs per month per topic, billed only as Apify compute usage, and the lightest tier keeps that compute low
- Sources filter accepts any subset of reddit, tiktok, instagram, youtube, hackernews, polymarket, github, threads, pinterest, and bluesky
- Residential proxy is on by default so scheduled runs are not blocked by platform 403 responses
- MCP endpoint lets an agent run a report and read the result in one tool call
| Tier | What it includes | When to use it |
|---|---|---|
| Quick Scan | Reddit, Hacker News, Polymarket, GitHub at base result count, optional AI synthesis (on by default) | Daily pulse checks on technical or financial topics where compute should stay minimal |
| Standard | Quick Scan sources plus TikTok, Instagram, YouTube at base result count, optional AI synthesis (on by default) | The default for most teams and for weekly monitoring of consumer or brand topics |
| Deep Intel | Same platforms as Standard with more results per source, plus comment threads and video transcripts | Quote mining, content research, and sales prep where you need what people said, not just headlines |
| Full Brief | Deep Intel depth plus AI web grounding via Perplexity for the synthesis | Executive briefings and launch reviews that need citations beyond the ten social sources |
How to do it
- 1.Open https://apify.com/openclawai/30days-trend-intel, enter a topic, and pick a timeframe of 7, 14, 30, or 90 days.
- 2.Choose a tier (Quick Scan, Standard, Deep Intel, or Full Brief), leave AI synthesis on, and optionally limit the sources list; from August 31, 2026 no tier carries a per-report fee.
- 3.Run it once to check the clusters and momentum label, then save the input as a schedule on Apify to build a weekly or daily series.
- 4.For agents, connect https://mcp.apify.com/?tools=fetch-actor-details,openclawai/30days-trend-intel and call the actor by name with the same input.
Questions, answered
How much does cross-platform trend monitoring cost with Datapika?
From August 31, 2026 the actor is free to run: there is no per-report charge for any tier, and you pay only for your own Apify platform usage, meaning the compute a run consumes plus any residential proxy bandwidth, which Apify bills at its standard rates. No platform API keys or subscriptions are needed. Deeper tiers such as Deep Intel and Full Brief fetch more per source and therefore use more compute than a Quick Scan, so pick the lightest tier that answers your question for high-frequency schedules.
Do Reddit, TikTok, and YouTube have official APIs for trend tracking?
Each platform has its own developer API with separate keys, quotas, and approval steps, and a self-built tracker breaks whenever one of them changes its terms or rate limits. Datapika's actor removes that setup: one input covers all sources, credentials are handled by the actor, and residential proxies are on by default. You still get engagement figures the platforms themselves report, such as upvotes, views, and likes.
Which platforms are included, and can I limit the report to a few of them?
The README lists 13+ platforms searched, and the sources filter exposes ten by name: Reddit, TikTok, Instagram, YouTube, Hacker News, Polymarket, GitHub, Threads, Pinterest, and Bluesky. Leave the filter empty to search everything your tier allows, or pass a subset such as tiktok and instagram for a consumer brand. Note that TikTok, Instagram, and YouTube require the Standard tier or above.
How fresh is the data in a trend report?
Every run fetches live data at execution time, so there is no cached index behind the results. The timeframe you pick (7, 14, 30, or 90 days) controls how far back each platform search reaches, and generated_at records the exact run time. For monitoring, a 7-day window on a weekly schedule gives non-overlapping slices, while a 30-day window smooths out single viral spikes.
What happens if the AI synthesis step fails?
You still receive the complete raw dataset: ranked candidates with engagement scores, cross-source clusters, and per-platform counts. The executive_summary, key_findings, and best_takes fields come back empty, or reduced to a plain extraction of the top-ranked titles, and from August 31, 2026 the run costs only the Apify compute it used. A pipeline should therefore treat the synthesis fields as optional and rely on the scored results as the source of truth.