Viral Platform Documentation

Performance stats

Baseline engagement numbers for your recent posts so every other card has context.

• Avg likes, avg comments, and avg reshares across analyzed posts.
• Engagement rate (likes relative to followers) and posts analyzed count.
• Follower count (editable if the scrape needs a correction).

Where you stand

A short LLM summary of how your habits compare to gold and bronze creators in your brackets — what to keep copying and what to drop.

• Requires bracket research so gold/bronze context exists.
• Highlights good practices vs. habits that look like bronze-level reach.

Social Selling Index

LinkedIn-only card. Pulls your Social Selling Index score and the four pillars so you can see how LinkedIn ranks your selling presence — separate from post engagement.

• Overall score out of 100, plus each pillar out of 25.
• Industry and network rank when LinkedIn provides them.
• Not included in PDF export.

Content mix & timing

Three sub-views that answer what you post, what you write about, and when you post.

• What you post: media mix (text, image, video, carousel, etc.).
• Topics you write about and type of content you post (themes from profile analysis).
• Consistency: your posts/week vs LinkedIn’s majority, cadence tier, and best days/times by engagement.

Your voice

A deterministic vocabulary fingerprint from your LinkedIn posts — frequent words, signature phrases, and typical post length. Content ideas and MCP drafting reuse this so copy sounds like you, not generic AI.

• Word cloud built from your scraped posts (stopwords stripped).
• Signature phrases (common two-word patterns).
• Available over MCP via get_user_vocabulary (includes a promptReady block).

Content analyser

Working vs. weak engagement patterns across your recent posts — the “what’s landing” vs. “what isn’t” read.

• Summary of overall content performance.
• What’s working: patterns tied to stronger engagement.
• What isn’t working: patterns tied to flat or weak posts.

3-P content strategy

Classifies your posts into Personality, Problem, and Proof — plus Consistency / Perseverance (posting cadence vs LinkedIn’s majority). The first three keep a LinkedIn feed human, useful, and credible; Consistency shows how often you show up compared with most users.

• Donut mix of Personality / Problem / Proof with pillar cards and advice.
• Perseverance tiers from posts/week: Outlier (1/day), Influencer (5/wk), Content focused (3/wk), Novice (1/wk), Under performing (1/mo).
• Audience resonance: whether comments match the intent of each pillar.
• Content half-life: how long engagement lasts per pillar (comment lifespan math).
• Posts to classify control (5–50) limits how many recent posts enter the content-mix run.

Audience insights

Who shows up in your comments, how they communicate, and how your brand reads from the outside.

• Who engages most: top engagers, lead-quality signals, CSV export.
• Audience personality (PCM): Process Communication Model distribution, per-type evidence, and a How we determined this walkthrough (posts ranked by engagement → classifier LLM → stored outputs).
• Your branding, as read: voice and positioning tags from profile analysis.

Account optimization

Full profile audit — headline, bio, skills, topics — plus how your content lines up with gold vs. bronze brackets.

• Paste or refresh bio/skills for algorithm alignment reads.
• What works well / what hurts / do this first action stack.
• Headline, bio, skills, and topics cards with concrete edits.
• Your content vs. the brackets: keep, stop, gold gaps, bronze overlap.

Bucket analysis

On Bracket research, Bucket analysis runs the per-bracket LLM pass (gold → silver → bronze). Configure creators per bracket (1–3) and posts per creator (10–50), then Start. Stop cancels an in-flight run.

• Status banners show when discovery, scraping, or analysis is still running.
• When creators are ready, open Patterns for the full playbook.

Creator grid & bracket averages

Each bracket shows a creator grid with followers, engagement, and post counts. Drag creators between brackets, open details, refresh posts, pick from the pool, remove, or replace by URL.

• Bracket averages strip: avg likes, comments, reshares, engagement rate, engagement score, follower range.
• Curating brackets changes what Patterns and synthesis optimize against.

Patterns playbook

After bucket analysis, Patterns is “Your bracket playbook”: start doing what gold does, fix silver gaps, stop bronze habits. Account-level audit stays under Your content → Account optimization.

• Viral only: analyze breakout posts per bracket (on) or include every post (off). Both modes are cached separately.
• Run research again to refresh the playbook after you change creators.
• Edit research prompts to customize the bracket analysis system prompt.

Start doing / Fix next / Stop doing

Three overview action cards from bronze → gold synthesis — the fastest way to act on bracket research.

• Start doing: gold habits to copy this week.
• Fix next: what silver does that still isn’t gold (gap + why not gold).
• Stop doing: bronze anchors that keep reach flat (habit + why it keeps you small).

Bronze → gold synthesis

Cross-bracket ladder summary that ties the three overview cards together. Expand nuances when you want the longer reasoning behind the short synthesis.

Per-bracket pattern cards

Each populated bracket (Gold / Silver / Bronze patterns) includes a viral profile, word overlap, dimension cards, and named pattern cards.

• What viral looks like: action summary, avg engagement, format signals, top viral post examples with hooks.
• Word overlap: shared vocabulary score, avg words per post, top overlapping words (computed, not LLM).
• Hooks, CTAs, Topics, Copy length, Angle, Target audience — dimension cards with action line, frequency, why it works, and examples.
• Topics can open Google Trends for rising interest in that niche language.
• Extra named pattern cards from the Brendan Kane–style analysis when the model surfaces them.

mcp.json example (Cursor / similar)

{
  "mcpServers": {
    "viral": {
      "url": "https://your-app.vercel.app/api/mcp/mcp",
      "headers": {
        "Authorization": "Bearer YOUR_TOKEN"
      }
    }
  }
}

Claude Code (CLI)

After generating a token in Viral Settings, add the server with the CLI command shown once in the connect dialog:

claude mcp add --transport http viral https://your-app.vercel.app/api/mcp/mcp --header "Authorization: Bearer YOUR_TOKEN"

Codex CLI example

export VIRAL_MCP_TOKEN="your-token-here"
codex mcp add viral --url https://your-app.vercel.app/api/mcp/mcp --bearer-token-env-var VIRAL_MCP_TOKEN

Hermes config.yaml example

mcp_servers:
  viral:
    url: "https://your-app.vercel.app/api/mcp/mcp"
    headers:
      Authorization: "Bearer YOUR_TOKEN"

Overview, ICP & voice

• get_mcp_agent_readme — operating instructions for AI agents (workflows, drafting rules, tool map).
• get_app_overview — index of available data across your account (includes icp + userVocabulary flags).
• get_business_context — full pack: business profile, ICP, brand style, LinkedIn analysis, research fingerprint.
• get_icp — Ideal Customer Profile only (title, industry, size, problem) plus audience/niche/themes.
• get_user_vocabulary — Your voice word cloud + signature phrases from their posts (promptReady block included).
• get_content_fingerprint — topic/positioning fingerprint used for lookalike research.
• update_business_context — update business profile, ICP, style, LinkedIn analysis notes, or format (subscription required).
• get_website_analysis — latest website crawl and analysis.
• crawl_website — crawl a site with Firecrawl (subscription required).
• analyze_website — LLM analysis of a stored crawl (subscription required).

LinkedIn profile & posts

• get_linkedin_profile_analysis — profile health and posting patterns.
• list_my_posts — your recent LinkedIn posts with metrics.
• refresh_linkedin_profile — re-scrape profile and posts (subscription required).

Creator research

• start_creator_research — start lookalike discovery + background ticks (subscription required).
• continue_creator_research — schedule another background tick without the UI open.
• get_creator_research_run — current lookalike research run metadata.
• list_creators — creators in gold/silver/bronze brackets.
• get_creator_posts — posts for a specific lookalike creator.
• get_bracket_analyses — per-bracket analysis summaries (includes shared vocabulary / word overlap).
• get_bracket_insights — exportable bracket insights bundle.
• get_bracket_synthesis — cross-bracket synthesis report.
• get_bracket_topic_trends — Google Trends for bracket topics.
• get_user_comparison — your metrics vs. bracket averages.
• compare_creators — head-to-head creator comparison (includes distinctive vocabulary).
• list_comparison_creators — creators available for comparison.
• get_research_prompts — saved research prompt templates.
• save_research_prompt — save a research prompt (subscription required).

Viral research & content-design import

• list_viral_research / get_viral_research / run_viral_research — stored viral research runs.
• import_viral_research — portable full research package for Codex/Claude/Hermes local storage.
• import_viral_research_ideas — themes, patterns, and ideas package.
• list_content_design_templates / get_content_design_skill — user design files + skill markdown.
• import_content_design_skill — pull (default) or push content-design.md skill.
• export_content_design_templates / import_content_design_templates — labeled templates/assets.
• get_meeting_content_skill / import_meeting_content_skill — platform Meeting ↔ Content skill (pull for all; push Jasper-only). Skip when user uses Fireflies/Fathom/other external meeting tools.
• get_podcast_content_skill / import_podcast_content_skill — platform Podcast → Content skill (pull for all; push Jasper-only; repo default podcast-2-content.md). Offer only if user makes podcasts.

Content ideas & drafts

• generate_content_ideas — generate new post ideas using ICP + Your voice (subscription required).
• get_generated_content_ideas — last generated idea batch.
• list_saved_content_ideas — your saved idea library.
• save_content_idea — save an idea (subscription required).
• unsave_content_idea — remove a saved idea (subscription required).
• simulate_draft — three-persona draft wind tunnel using ICP for Ideal Client (subscription required).

3-P strategy, resonance & lifespan

• get_three_p_content — stored Personal / Problem / Proof mix.
• run_three_p_content_analysis — classify posts into 3-P pillars (subscription required).
• get_audience_resonance — stored comment-vs-pillar resonance.
• run_audience_resonance_analysis — run resonance analysis (subscription required).
• get_content_lifespan — algorithmic half-life per pillar.
• refresh_content_lifespan — refresh comments/snapshots and recalc half-life.

Audience & optimization

• get_top_engagers — cached top engager analysis.
• run_top_engagers_analysis — run fresh engager analysis (subscription required).
• export_top_engagers_csv — CSV export of top engagers.
• get_social_selling_index — LinkedIn Social Selling Index score.
• refresh_social_selling_index — refresh SSI (subscription required).
• get_pcm_audience — Process Communication Model audience read.
• run_pcm_audience_analysis — run PCM analysis (subscription required).
• get_account_optimization — profile optimization recommendations.
• run_account_optimization — run optimization audit (subscription required).

Trends & content comparison

• get_topic_trends — latest topic trends search results.
• start_topic_trends_search — start a trends search (subscription required).
• check_topic_trends_search — poll an in-progress trends search.
• list_content_piece_comparisons — saved content piece comparisons.
• compare_content_pieces — compare posts or drafts (subscription required).

Recommended assistant flow

• 1. get_app_overview — confirm profile, ICP, and vocabulary flags.
• 2. get_bracket_synthesis + get_user_comparison — gold best practices and bronze negative patterns.
• 3. get_icp + get_business_context + get_user_vocabulary — buyer, company, voice.
• 4. Use the user’s idea or generate_content_ideas / get_generated_content_ideas.
• 5. Generate the visual first (infographic or slider), then draft the post in their register.
• 6. simulate_draft — iterate until readiness looks solid.

Install paths

• Claude.ai: /skills/viralz-content-creation-for-claude.txt → Project → Files (do not upload .md or .zip)
• Zip (Cursor / Claude Code / Hermes): /skills/viralz-content-creation.zip
• Project skill folder: .agents/skills/viralz-content-creation/ (after unzip)
• Public browse: /skills

Custom business packages (admin)

Admins can build a per-business skill package in Settings → Admin → Business content skills: SKILL.md, slider/infographic background files, signature and CTA templates, HTML examples, and up to 9 LinkedIn creator URLs.

Creator URLs are staged on the package only — they are not scraped or added to bracket research until you intentionally apply them later. Build the package first, then Connect to user by email when the account exists.

Assigned users download the full zip (or individual template assets) from Settings → AI assistants.

Verify MCP health

Check the public status page at /status for live MCP, API, and database health. The MCP endpoint URL and HTTPS status are listed at the bottom of the status page.