README.md

metadata

title: MCPilot
emoji: 👁
colorFrom: yellow
colorTo: yellow
sdk: gradio
sdk_version: 6.0.1
app_file: app.py
pinned: false
license: apache-2.0
short_description: Your MCP Server for API Integration Testing
tags:
  - mcp-in-action-track-enterprise

🚀 MCPilot

Your MCP Server for API Integration Testing

🎥 Demo Video

Watch the demo video to see how MCPilot works. I also wrote about MCPilot on X too

💰 Save $50-200 per project on LLM token costs during development
⚡ Test 6 APIs instantly - GitHub, Slack, Linear, Notion, Stripe, Sentry
🎭 Mock mode = 100% FREE - Unlimited testing without spending a cent

✨ MCP-Native • 🎭 Mock & Real Modes • 💰 Token Cost Tracking • 🔐 Permission Validation

---

💰 Why Developers Love MCPilot

The Hidden Cost Problem

When building AI agents with API integrations, developers unknowingly burn through money:

• 💸 $2-5 per testing session on LLM tokens
• 💸 $50-200 per project during development
• 💸 10-100x overspending by not comparing LLM providers

The MCPilot Solution

Mock Mode (Development)

• ✅ Test unlimited API calls for FREE
• ✅ Zero LLM token costs
• ✅ Instant responses, no rate limits
• ✅ 100% savings during development

Cost Intelligence (Before Production)

• 🎯 Estimate exact token costs
• 🎯 Compare 8 LLM providers
• 🎯 Find 10-100x cheaper options
• 🎯 Know costs BEFORE deploying

Result: Develop fearlessly in mock mode, deploy confidently knowing exact costs!

---

🎯 What is MCPilot?

MCPilot is a Model Context Protocol (MCP) server that enables instant API integration testing without setup, credentials, or costs. Built specifically for AI agents and MCP workflows, it provides:

• 6 Service Integrations: GitHub, Slack, Linear, Notion, Stripe, Sentry
• 35+ API Actions: Complete CRUD operations for each service
• Dual Modes: Mock (free, instant) and Real (production validation)
• Cost Intelligence: Accurate token estimation across 8 LLM providers
• Permission Guardian: Validate credentials before production deployment

🌟 Why MCPilot?

The Problem: Wasting Money on Development

Building AI agents with API integrations burns money unnecessarily:

• ❌ LLM token costs during testing ($2-5 per session)
• ❌ API costs for development calls ($10-50 per project)
• ❌ Wrong LLM choice (paying 100x more than necessary)
• ❌ Permission errors caught in production (wasted deployment costs)
• ❌ No cost estimation before committing to an LLM

Typical waste: $50-200 per project, just during development phase.

The Solution: MCPilot Eliminates Waste

MCPilot lets you:

• ✅ Test for FREE with realistic mocks (100% savings)
• ✅ Estimate costs across 8 LLM providers (find 10-100x savings)
• ✅ Validate permissions before deployment (avoid wasted deploys)
• ✅ Compare providers to pick the cheapest option
• ✅ Switch to real when ready, with one flag

Result: Zero wasted money during development, informed decisions for production.

📚 Core Features

🎭 Mock Mode (Default)

• Zero setup required
• Instant responses
• Realistic data from official API specs
• Perfect for development and testing

🌐 Real Mode

• Toggle with one flag
• Actual API calls
• Production validation
• Cost tracking

🔐 Permission Validation

• Check credentials before deployment
• Detailed scope requirements
• Format validation for all services
• Helpful error messages with docs links

💰 Token Cost Tracking

• Accurate counting with tiktoken
• Support for 8 LLM providers
• Cost comparison across models
• Session analytics and savings reports

🔍 Mock vs Real Comparison

• Validate mock accuracy
• Detect API schema changes
• Similarity scoring
• Actionable recommendations

🛠️ Supported Services

Service	Actions	Mock	Real	Status
GitHub	5	✅	✅	Complete
Slack	6	✅	🚧	Mock ready
Linear	6	✅	🚧	Mock ready
Notion	6	✅	🚧	Mock ready
Stripe	6	✅	🚧	Mock ready
Sentry	6	✅	🚧	Mock ready

Total: 35+ API actions available

💡 Use Cases

AI Agent Development

# Test your Claude agent's GitHub integration
agent_prompt = "Review PR #123 and summarize changes"
# MCPilot provides instant GitHub data without setup

Cost Optimization

# Compare token costs before choosing an LLM
comparison = compare_costs(1000, 500)
# See costs across Claude, GPT-4, Gemini, etc.

Pre-Production Validation

# Validate permissions before deployment
validate_permissions('github', 'create_pull_request', credentials)
# Catch permission issues early

Integration Testing

# Test multi-service workflows instantly
get_sentry_error() → create_linear_issue() → post_slack_alert()
# All without any API setup

📊 LLM Provider Support

Track token costs across:

• Claude: Sonnet 4.5, Opus 4, Haiku 4
• OpenAI: GPT-4 Turbo, GPT-4o, GPT-3.5 Turbo
• Google: Gemini Pro, Gemini Flash

Pricing updated: 2025-11-30

🎨 Web Interface

MCPilot includes a beautiful Gradio web interface with 6 tabs:

1. 🎭 Simulate Integration - Test API calls
2. 🔐 Validate Permissions - Check credentials
3. 📋 List Actions - Discover capabilities
4. 💰 Compare Costs - LLM pricing comparison
5. 🔍 Mock vs Real - Accuracy validation
6. 📊 Session Report - Analytics & savings

Built with ❤️ for the AI agent community

🚀 MCPilot - Navigate API integrations with confidence