MCP integration

Add Hooklayer to your agent in 30 seconds.

One config block. 12 viral-intelligence tools instantly callable from Claude Desktop, Cursor, n8n, or any HTTP MCP client. The recommended_chain field on analyze_account is advisory data — an ordered list of possible next tool calls with pre-filled parameters and reasons. The user or assistant decides whether to invoke them; nothing is auto-executed.

25 free credits · no card · works on free Claude Desktop with mcp-remote bridge

v2The evidence layer is live. Every score now ships with the signals that produced it, source provenance, and a counterfactual. See what changed →
Step 1

Pick your MCP client.

~/Library/Application Support/Claude/claude_desktop_config.json (macOS) %APPDATA%\Claude\claude_desktop_config.json (Windows)
{
  "mcpServers": {
    "hooklayer": {
      "command": "npx",
      "args": ["-y", "hooklayer", "hl_live_..."]
    }
  }
}
Steps
  1. 1Open the config file above (create it if missing)
  2. 2Paste the JSON, replace hl_live_... with your real key
  3. 3Quit Claude Desktop completely, then reopen it
  4. 4Look for "hooklayer" in the 🔌 connector list — should show 14 tools
Full MCP docs
Step 2 — see the advisory chain

One call in. A ranked list of what to try next.

Watch Claude Desktop run analyze_account(@mrbeast) and read the advisory recommended_chain field — a ranked list of possible next tool calls (like match_voice, trend_pulse, and viral_remix) with pre-filled parameters and reasons. The user or assistant chooses which ones to actually invoke.

Don't want to watch? Try it live in the playground — no signup required.

Step 3 — what your agent gets

Twelve tools, one Bearer token.

analyze_account
5cr

Deep TikTok, YouTube, or Instagram creator analysis with viral DNA scores, format fingerprint, and an advisory recommended_chain field the assistant can choose to follow up with.

score_hook
1cr

Score any hook 0-100 against proven viral patterns. Returns 3 stronger rewrites + matched pattern. The first-call sanity check.

viral_remix
3cr

URL or transcript → fresh script mirroring the viral DNA. Scene-by-scene with camera shots and text overlays.

trend_pulse
1cr

Real-time rising opportunities + saturated patterns per niche. Format / hook / style trends, not news topics.

find_viral_template
1cr

Niche-fit ranked templates with hook pattern, format structure, and example URLs.

search_videos
2cr

Keyword search across TikTok / YouTube / Instagram ranked by weighted engagement. Returns URLs + hook patterns for follow-up analysis.

match_voice
2cr

Extract voice DNA from 3+ samples, rewrite a draft in that voice. Returns reusable prompt instructions.

predict_virality
2cr

Score a draft script for viral potential before publishing. Returns viral DNA, retention diagnosis, share trigger.

brief_to_blueprint
7cr

Turn a client brief into a ready-to-ship blueprint. Chains templates + trends + hook_score + predict internally in one call.

watch_account
0 or 5cr

Track a creator so future calls can detect what changed. 0 credits when a recent analysis can be reused for the baseline; otherwise 5.

list_watches
0cr

List every creator the authenticated user is tracking. Pure read-only DB query; never charges credits; never modifies state.

get_changes
5cr

Detect what meaningfully changed since the previous saved baseline for a tracked creator. Appends a new snapshot; returns high/medium shifts only.

product_scout
5cr

TikTok Shop product research. Scans a niche and returns ranked products with provider evidence plus a PROMOTE / WATCH / SKIP verdict carrying the reason. Deterministic scoring, no LLM in the ranking path. PRO and AGENCY plans.

analyze_product
10cr

Full teardown of one TikTok Shop product: buyer language from real reviews, the objections worth answering on camera, and the creative next step. Never returns revenue or GMV. PRO and AGENCY plans.

Free tier. No card. Live in your agent in 5 minutes.

25 lifetime credits at signup. Same SKU works for REST and MCP. No seat charges, no AI Maker tax.