5M+
Ads indexed
Find winning videoson IG and TikTokKeep your list fresh
A discovery workflow for overseas short-video research: mine keyword pools, search for TikTok video candidates, dedupe your benchmark list, and summarize each scan run in one place.
2-4 min
To a shortlist
480+
Teams using HiClaw
Used by brands, DTC teams, and creators
Get a shortlist of latest proven ads, then turn into hooks, scripts, and variations.
Why this skill exists
Viral research often starts as scattered search tabs and half-maintained spreadsheets. This pipeline turns that into a repeatable discovery flow with canonical artifacts on disk and a clear hand-off into deeper video analysis after the shortlist is ready.
TikTok-onlyby design
The current pipeline explicitly discovers TikTok video URLs. It does not quietly pretend to cover Shorts or Reels before the tooling actually supports them.
Artifacts staycanonical
Keywords, benchmarks, accounts, and dated logs remain the operational source of truth. The final report summarizes the run but does not replace the actual files.
Search-basedand auditable
Discovery keeps query intent and evidence, so benchmark additions are explainable and easier to maintain across repeated scans.
What the pipeline produces
The page mirrors the skill spec: discovery first, list maintenance on disk, and a final delivery report for the current run. This is benchmark discovery, not script generation.
1 · `keywords.json`
Grouped intents, short-tail and long-tail keywords, plus executable query templates for future scans.
2 · `benchmarks.csv`
TikTok video URL candidates with dedupe, canonical URLs, picked reason, and discovery evidence for auditability.
3 · `accounts.csv`
Account-level rollup created from the benchmark list so repeat discoveries stay organized by creator.
4 · Run log
Optional dated scan summary under `out/viral-discovery/runs/` for recurring discovery workflows.
5 · Delivery report
A chat-facing `report.md` for the current run that summarizes scope, artifacts updated, and notable additions.
6 · Raw search trail
Search evidence can be kept alongside the pipeline outputs so the shortlist remains explainable, not hand-wavy.
Conversation flow
Minimal intake, then run the right path: initialization, scan/update, or keywords only. Discovery stays grounded in the A → B → C → D flow from the skill.
1
Intake
Share the niche, target market, language, and how many candidates you want. If data already exists, say whether this is initialization or a scan/update run.
2
Mine keywords
The pipeline builds grouped search intents first when needed, so discovery is grounded in reusable keyword sets instead of ad hoc searching.
3
Discover TikTok URLs
Search-based discovery finds TikTok video candidates, keeps the evidence trail, and merges results into the benchmark list by canonical URL.
4
Dedupe and report
The run updates benchmarks, rolls up accounts, optionally writes a dated log, then produces a final report for what changed in this run.
Input cheat sheet
- Core input: niche, market/country, language, and target candidate count
- Optional: seed keywords, existing benchmark/account files, or whether this is a recurring scan
- Output mode: initialization, scan/update, or keywords-only run
- Scope guardrail: TikTok discovery today; other short-video platforms are not part of this pipeline yet
Related use cases
Browse more in the skills directory
Viral video teardown after you shortlist links
Keyword pools for ongoing TikTok scans
Benchmark list maintenance across repeated runs
Creator/account rollups from discovered URLs
Search-based discovery with audit trails
Run summaries for weekly or daily monitoring
Niche, market, and language-specific scans
Brave-powered discovery before deep analysis
TikTok-only scope with explicit guardrails
Hand-off into downstream video analysis
Ready to build your benchmark list?
Open WebChat with this skill context, share the niche and market you want to scan, and get a discovery run that updates keywords, benchmarks, accounts, and the delivery report.