Hooklayer
All tools/analyze_account5 creditsresearch

What is analyze_account?

analyze_account is the Hooklayer MCP tool that takes one TikTok handle and returns a creator's full viral DNA — viral / replicability / originality scores 0-100, format fingerprint (cadence, duration, hook reuse rate), top 5 videos with transcripts, content gaps, and a recommended_chain field pre-filling the next 3 tool calls your agent should run.

Get free API keyTry in playground

The flagship Hooklayer tool. One call in, four calls out — analyze_account returns deep intelligence plus a `recommended_chain` array that names the next tools (match_voice, trend_pulse, viral_remix) with parameters pre-filled. Designed so AI agents read the chain field and continue the workflow without prompt engineering.

Every score ships with cited evidence. The viral_dna_signals, replicability_signals, and originality_signals arrays each contain 3-5 named signals with evidence strings quoting concrete facts from the videos. Agents cite the phrase, not paraphrase the number.

Costs 5 credits. 1-hour cache per handle. TikTok only in v1 (Instagram and YouTube ship in v2). Returns a `would_fail_because` counterfactual naming the non-portable element (face/voice/credibility) that blocks replication for a no-name account.

Inputs & outputs

Endpoint: POST /api/v1/account/analyze

Inputs

  • handlestringrequired

    TikTok handle with or without @ prefix

  • platformstring

    Currently only "tiktok" — Instagram and YouTube ship in v2

Output fields

  • profile

    Username, follower count, avg views, niche, region

  • viral_dna

    viral_dna_score, replicability_score, originality_score, consistency_score, hook_reuse_rate, audience_fatigue, plus 3 signals[] arrays with cited evidence

  • format_fingerprint

    avg_duration_sec, words_per_second, pacing, sound strategy

  • recent_videos

    Top 5 videos with views, transcripts, hashtags, source_url, posted_at

  • steal_map

    3 transferable patterns with concrete "how" + applicability + example_remix

  • recommended_chain

    3 pre-filled next-tool calls with action_class, confidence, cost, expected_output

  • provenance

    data_source, fetched_at, data_age_days, videos_analyzed

  • quality

    level (full | partial | degraded) + reason

cURL

curl -X POST https://hooklayer.dev/api/v1/account/analyze \
  -H "Authorization: Bearer hl_live_..." \
  -H "Content-Type: application/json" \
  -d '{"handle": "@humphreytalks", "platform": "tiktok"}'

Example prompts

Paste any of these into Claude Desktop (with Hooklayer connected) to see the live response.

Creator teardown with cited evidence

Use Hooklayer to analyze @humphreytalks on TikTok. Show me the full viral_dna block including viral_dna_signals, replicability_signals, and originality_signals arrays — I want the evidence quotes verbatim, not paraphrases. Also surface the would_fail_because field and provenance.video_post_dates so I can confirm the data is recent.

Expected output: Returns Humphrey's DNA scores (viral 87 / replicability 85 / originality ~78), 3-5 signals per DNA dimension with cited evidence strings, the would_fail_because counterfactual, and the recommended_chain with confidence/cost/action_class per step.

Steal-map for niche transfer

Analyze @herfirst100k and show me the steal_map — I need 3 transferable patterns with concrete how-to lines and example_remix for each. Then explain which patterns are face-dependent vs format-dependent so I know which ones I can actually copy.

Expected output: Returns 3 patterns with what / how / applicability / example_remix, plus a replicability score that explains the non-portable formula problem (high credentials = lower portability).

Agentic chain auto-fire

Analyze @mrbeast and then execute the recommended_chain it returns — fire match_voice, trend_pulse, and viral_remix in order using the parameters Hooklayer pre-filled. Report what each tool returned and confirm the chain executed end-to-end.

Expected output: Demonstrates the agentic pattern. analyze_account returns 3 next-tool calls with params; the agent fires them sequentially; you see 4 tool outputs in one conversation.

Frequently asked

How is analyze_account different from a regular TikTok scraper?

A scraper returns raw stats. analyze_account returns scored intelligence — viral DNA breakdown, replicability scoring (face-dependent vs format-dependent), a steal_map of transferable patterns, and a recommended_chain telling your agent which tools to call next. Same data underneath, plus a scoring + chaining layer designed for AI workflows.

Why does it cost 5 credits when score_hook is only 1?

analyze_account triggers: TikTok profile fetch + top 5 video fetch + transcript extraction (Whisper) for each + pattern analysis + a meta-analysis Claude call producing the DNA scores and recommended_chain. Net AI cost is roughly 5× a single score_hook call. The 1-hour cache makes follow-up calls on the same handle free at the upstream layer.

Does the 1-hour cache mean I get charged twice if I run it twice within an hour?

Yes — caching is upstream (data fetching) only. The Hooklayer credit ledger charges per API call regardless of cache hits. This keeps cost predictable for agent loops and avoids penalizing iterative workflows. Set a 24-hour determinism cache hash on the meta-analysis ensures the same handle returns identical scores within that window.

How does the recommended_chain decide what tools to call next?

A Claude Sonnet meta-analysis layer reads the creator's signals — consistency score, hook reuse rate, replicability — and picks 3 next-tool calls with confidence levels (high / medium / low) derived from those deterministic signals. High consistency + ≥3 transcripts → match_voice high. High replicability → viral_remix high. The pattern is auditable in the response.

What's the action_class field for?

Authority taxonomy. Each chain step carries action_class: research (read-only, lowest authority), synthesize (generate in-memory), draft, publish, or account. All 7 current Hooklayer tools are research or synthesize. The field lets agents and reviewers filter steps by authority before executing — answers "what kind of power does this exercise" separately from "does it mutate state."

Why TikTok only? Will Instagram and YouTube be supported?

TikTok's API surface and viral mechanics are sufficiently different from Instagram Reels and YouTube Shorts that a single rubric would dilute scoring quality. Each platform gets its own analyze_account variant in v2 with platform-specific signals (Reels remix culture, Shorts watch-completion bias). Roadmap target: Q3 2026.

Related tools

viral_remix
URL or transcript → fresh script mirroring the viral DNA. Returns scene-by-scene structure, camera shots, text overlays, and a verify_hook block.
match_voice
Extract voice DNA from reference samples and rewrite drafts in that voice. Returns voice_metrics (TTR, filler rate, signature phrases) plus a reusable prompt.
trend_pulse
Real-time rising opportunities + saturated patterns per niche. Aggregates Google Trends, YouTube velocity, Reddit hot, and internal corpus.

Try analyze_account in 30 seconds.

100 free credits at signup. No card. Works in Claude Desktop, Cursor, n8n, or any MCP client.

Get API keyInstall for Claude / Cursor / n8n