AgentStack
MCP verified MIT Self-run

Reactor

mcp-keugenek-reactor · by keugenek

Predict LinkedIn post performance before you publish. GO/WAIT/IMPROVE/NO verdict, impressions, debate score, cringe detection. CLI + MCP + AI Skill. Zero LLM calls.

No reviews yet
0 installs
7 views
0.0% view→install

Install

$ agentstack add mcp-keugenek-reactor

✓ scanned · ✓ verified — works with Claude Code, Cursor, and more.

Security review

✓ Passed

No issues found. Passed automated security review. · v0.1.0 How review works →

  • Prompt-injection patterns
  • Secret / credential exfiltration
  • Dangerous shell & filesystem operations
  • Untrusted network calls
  • Known-malicious package signatures

What it can access

  • Network access No
  • Filesystem access No
  • Shell / process execution No
  • Environment & secrets No
  • Dynamic code execution No

From automated source analysis of v0.1.0. “Used” means the capability is present in the source — more access means more to trust, not that it’s unsafe.

Are you the author of Reactor? Claim this listing to set pricing, connect Stripe payouts, and keep 70% of every sale.
Sign up to claim

About

🔮 Reactor — LinkedIn Post Analyzer & Engagement Predictor

[](https://npmjs.com/package/@keugenek/reactor) [](LICENSE) [](https://github.com/keugenek/reactor)

LinkedIn Post Analyzer — predict engagement before you publish. Get a clear GO/WAIT/IMPROVE verdict, estimated impressions, hook score, debate potential, and cringe detection.

Try it free: keugenek.github.io/reactor

npx @keugenek/reactor "Just shipped my side project. Here's what I learned."
  ✅ GO  14,772 impressions

  Debate    60/100
  Hook      0.34
  Cringe    8/100
  Auth      93/100
  Timing    1.38x

Why Reactor?

| Competitors | Reactor | |-------------|---------| | Give you a score (0-100) | Give you a decision: GO/WAIT/IMPROVE | | "Your hook is weak" | "You'll get ~2,500 impressions" | | Generic emotional analysis | Debate score + Cringe detection | | Web-only | CLI + MCP + Claude Desktop + Web | | Black-box AI | Transparent scoring rules |

Unique Metrics

  • Debate Score — Will people argue in comments? (controversy = engagement)
  • Cringe Detection — Is it too salesy, cliché, or inauthentic?
  • Estimated Impressions — Not just "good/bad" — actual reach prediction
  • Timing Multiplier — Best posting times for your audience

Quick Start

Web (Free)

Analyze your post → — no sign-up required

CLI

# Run without installing
npx @keugenek/reactor "Your LinkedIn post text"

# Or install globally
npm install -g @keugenek/reactor
reactor "Your post"

MCP Server (Claude Desktop / Cursor / Windsurf / Claude Code)

Add to your MCP config:

{
  "mcpServers": {
    "reactor": {
      "command": "npx",
      "args": ["@keugenek/reactor", "--mcp"],
      "env": {
        "REACTOR_API_KEY": "rk_free_your_key_here"
      }
    }
  }
}

Your AI assistant gets these tools:

| Tool | Description | |------|-------------| | predict_post | Full prediction: verdict + impressions + improvements | | analyze_hook | Hook analysis (curiosity, emotion, controversy) | | best_posting_times | Optimal posting windows for your audience | | reactor_register | Get free API key | | reactor_join_waitlist | Join Pro waitlist |

OpenClaw / AI Agent Skill

Copy skill/SKILL.md to your agent's skills directory. Auto-triggers on LinkedIn post discussions.

CLI Commands

reactor "Your post"                        # predict
reactor "Post" --type opinion_hot_take     # with content type
reactor "Post" --hour 9                    # with posting time
reactor --types                            # list 18 content types
reactor --discover                         # API features + tiers
reactor --register me@email.com            # get free API key (5/day)
reactor --waitlist me@email.com            # join Pro waitlist (50% off)

Content Types

Each type has a viral multiplier based on LinkedIn data:

| Type | Multiplier | Example | |------|-----------|---------| | technical_achievement | 9.72x | "I built X, here's how" | | opinion_hot_take | 5.00x | "Unpopular opinion: Y is overrated" | | research_publication | 1.28x | Original data, studies | | product_launch | 0.60x | "Excited to announce Z" | | rant | 0.15x | Pure complaints |

Run reactor --types for all 18 types.

API

Base URL: https://reactor-api.keugenek.workers.dev

| Endpoint | Method | Description | |----------|--------|-------------| | /v1/predict | POST | Predict post performance | | /v1/analyze-hook | POST | Analyze hook quality | | /v1/best-times | GET | Optimal posting times | | /v1/content-types | GET | List content types + multipliers | | /v1/register | POST | Get free API key | | /v1/discover | GET | Progressive discovery (all endpoints) |

Pricing

| Tier | Price | Features | |------|-------|----------| | Demo | Free | 2 predictions/day, basic verdict | | Free | Free | 5 predictions/day, hook analysis, best times | | Pro | £19/mo | Unlimited, custom personas, media analysis, A/B testing |

SEO Keywords

LinkedIn post analyzer, LinkedIn engagement predictor, LinkedIn post checker, LinkedIn virality checker, LinkedIn hook analyzer, LinkedIn post score, LinkedIn content analyzer, LinkedIn algorithm tool, LinkedIn post optimizer, LinkedIn engagement calculator, LinkedIn post performance predictor, LinkedIn cringe detector, LinkedIn debate score.

Privacy

  • We store only your email and usage count
  • No post content is stored or logged
  • All scoring is stateless
  • Delete your account anytime: reactor --delete

Architecture

┌─────────────┐     ┌──────────────────────┐
│  CLI / MCP  │────▶│  Cloudflare Worker   │
│  (this pkg) │◀────│  (scoring engine)    │
└─────────────┘     │  KV: rate limits     │
                    │  KV: registrations   │
                    └──────────────────────┘

Contributing

Issues and PRs welcome. The scoring engine is proprietary (Cloudflare Worker).


If Reactor saved you from posting a 💀, star the repo

Source & license

This open-source MCP server is cataloged on AgentStack and links to its original source — we do not rehost the code.

Install and usage instructions live in the source repository linked above.

Reviews

No reviews yet — be the first.

Versions

  • v0.1.0 Imported from the upstream source.