apify-mcpc

apify-mcpc plugin

Claude Code plugin for finding, evaluating, and running Apify Actors using the mcpc CLI.

Table of contents

• What it does
• Prerequisites
• Installation
• Skills
• Use cases
• Skill workflow
• Available mcpc tools
• Connection modes
• Cost tracking
• User-Agent tracking
• Plugin structure
• Features
• Testing
• License

What it does

• Searches Apify Store for the right Actor for any scraping/automation task
• Compares Actors by stats, ratings, and maintenance status
• Reads input schemas to build correct Actor inputs
• Runs Actors and retrieves structured results (sync or async with resume)
• Tracks run costs automatically via PostToolUse hook → ~/.apify-costs.log
• Covers 8 marketing intelligence use cases with domain-specific guidance, plus multi-actor workflow patterns

Prerequisites

• mcpc CLI: npm install -g @apify/mcpc
• Apify account with OAuth: mcpc mcp.apify.com login

Installation

From marketplace

# 1. Add the marketplace (once)
claude plugin marketplace add chocholous/apify-mcpc

# 2. Install the plugin
claude plugin install apify-mcpc@apify-mcpc

# 3. Verify
claude plugin list

Local development

claude --plugin-dir .

Skills

This plugin provides two slash commands:

Command	Description
/apify-mcpc	Main skill — find, evaluate, and run Apify Actors (full 7-step workflow)
/apify-status	Show active mcpc sessions and run cost history from ~/.apify-costs.log

Use cases

The plugin includes 8 use-case files with suggested Actors, search keywords, and gotchas, plus a multi-actor workflow patterns reference:

Use Case	Description
Audience Analysis	Demographics, follower behavior, engagement quality
Brand Monitoring	Reviews, ratings, sentiment, brand mentions
Competitor Intelligence	Competitor strategies, ads, pricing, positioning
Content Analytics	Engagement metrics, campaign ROI, content performance
Influencer Discovery	Find influencers, verify authenticity, partnerships
Lead Generation	B2B/B2C leads, contact enrichment, prospecting
Market Research	Market conditions, pricing, geographic opportunities
Trend Analysis	Emerging trends, viral content, content strategy
Multi-Actor Workflows	Chaining Actors, handle discovery via SERP, enrichment

Actor tables in use-case files are suggestions, not closed lists. The agent always searches Apify Store for the latest options.

Skill workflow

The skill enforces a 7-step workflow with mandatory gates to prevent wasted money on incorrect Actor runs:

1. Search — Find candidate Actors (always at least 2 searches: broad then narrow)
2. Fetch details — Read README, input schema, stats, ratings
3. Build input — Cross-reference schema with README for correct values
4. Validate — Challenge assumptions about identifiers, verify against target service
5. Run — Execute the Actor (test run with limit=1 first when uncertain)
6. Verify — Sanity-check results before presenting (zero results = wrong input, not "no data")
7. Get results — Fetch full dataset with field selection and pagination

Two mandatory gates stop the agent before running:

• After Step 2: Show user the Actor choice + planned input, wait for confirmation
• After Step 4: If any input value is guessed, tell user before running

The agent determines scope first — docs-only (Crawlee/platform questions), find-only (Actor comparison without running), or full pipeline (data extraction) — to avoid unnecessary Actor searches or runs.

Available mcpc tools

Session	Tool	Purpose
@apify	search-actors	Search Apify Store by keywords
@apify	fetch-actor-details	Get README, input schema, stats, pricing
@apify	call-actor	Run an Actor (sync or async)
@apify	get-actor-output	Fetch dataset results with field selection
@apify	get-actor-run	Check run status and stats
@apify	apify-slash-rag-web-browser	Quick web search / URL fetch without Actor workflow
@apify-docs	search-apify-docs	Search Apify/Crawlee documentation
@apify-docs	fetch-apify-docs	Fetch a specific documentation page

Connection modes

• Full mode (@apify) — requires OAuth (mcpc mcp.apify.com login), provides all tools including Actor execution
• Docs-only mode (@apify-docs) — no authentication needed, only documentation search and fetch