| name | prometheus-cli |
| preloaded | true |
| description | Search TikTok viral videos, App Store rankings, hook analysis, app strategy, and content research via SGE Prometheus MCP / Passistant. Requires SGE_API_KEY. Use when researching viral video hooks, benchmarking app rankings, analyzing content trends, pulling App Store review data, or when a prompt mentions "prometheus", "Passistant", "SGE", "Social Growth Engineers", "viral content", "app store rankings", "TikTok research", or "hook analysis". |
Prometheus — SGE Viral Content Intelligence
Search 50,000+ curated TikTok & Reels viral videos, app marketing profiles, and strategy articles. Live App Store rankings, reviews, and revenue estimates. Hook analysis and trend detection — all via MCP.
Homepage: https://www.socialgrowthengineers.com/prometheus
Requires: SGE_API_KEY environment variable.
Setup
Claude Code
claude mcp add --transport http --header "Authorization: Bearer $SGE_API_KEY" prometheus https://www.socialgrowthengineers.com/api/mcp
Or add to your project's .mcp.json:
{
"mcpServers": {
"prometheus": {
"url": "https://www.socialgrowthengineers.com/api/mcp",
"headers": {
"Authorization": "Bearer ${SGE_API_KEY}"
}
}
}
}
Cursor / OpenClaw / Other MCP Clients
Same config — point your MCP client at the URL above with your API key as Bearer token.
export SGE_API_KEY=sk_mcp_your_key_here
Get your API key at socialgrowthengineers.com/settings (requires SGE subscription).
Tool Discovery
All tools are auto-discovered via tools/list — no need to memorize them. When connected, your agent automatically sees every available tool with full parameter schemas.
In Hermes or cron contexts, call Prometheus through mcporter even if the prometheus skill alias was not loaded:
mcporter list prometheus
mcporter call prometheus.hook_intelligence scope_type=vertical scope_primary_vertical=finance --output json > hook-intel.json
mcporter call prometheus.search_videos query="investing tips" platform=tiktok limit:20 --output json > videos-investing.json
mcporter call prometheus.ast_get_rankings category_id=6015 limit:25 --output json > rankings.json
mcporter output may be JSON-like MCP text content rather than strict JSON for some tools. If jq fails, inspect the raw file and extract the content[].text payload or quoted text: blocks before synthesis. Keep the raw file regardless.
Tools fall into five categories:
| Category | What they do |
|---|
| Search & Discovery | Search articles, viral videos, and app profiles across SGE's curated database |
| Similarity | Find similar articles, apps, or videos using semantic embeddings |
| Live App Store | Real-time iOS App Store data — search, details, rankings, reviews, revenue estimates |
| Hook Intelligence & Strategy | Analyze winning hooks, compare strategies, detect trends across verticals |
| Video Analysis & Reports | Analyze any TikTok/Reels URL with AI vision, generate shareable research reports |
Usage Patterns
Research a vertical
1. search_videos -> "fitness hooks" (platform: tiktok, limit: 20)
2. hook_intelligence -> scope_type: "vertical", scope_primary_vertical: "fitness"
3. generate_report -> "fitness app viral marketing"
Competitive analysis
1. search_apps -> "meditation" (find App Store IDs)
2. strategy_compare -> left: Calm, right: Headspace
3. ast_get_app_details -> deep dive on each
4. ast_get_reviews -> sentiment analysis from users
Discover trending hooks
1. hook_trends -> scope_type: "market", current vs prior 30-day windows
2. search_videos -> top rising hook type (minViews: 1000000)
3. get_similar_videos -> find more like the best performer
Analyze a competitor's video
1. analyze_video -> paste their TikTok URL
2. get_similar_videos -> find SGE videos with similar hooks
3. hook_intelligence -> scope_type: "app", scope_app_name: "competitor"
Slack Bot
The same tools are available via the Prometheus Slack Bot — plus memory, scheduled reports, and app tracking.
Install from your SGE settings page.
Attribution warning
- Never attribute "Bloom: Learn to Invest" to our Bloom. That's a competitor app on the App Store. Our Bloom's identifiers are distinct — confirm the bundle ID / product ID when citing App Store data or results returned by Prometheus.