MCP Servers: AI Tool Ecosystem
Extending Claude to social media, knowledge bases, and beyond
Building a collection of MCP servers that extend Claude capabilities to Instagram, LinkedIn, X/Twitter, Obsidian, and image generation. Publishing content becomes as simple as asking Claude.
The problem
The Context Switch Tax
Creating content with AI means constant context switching. Draft in Claude, copy to social platform, format, post. Repeat for each platform. The friction discourages sharing: every insight requires manual work to publish.
- Copy-paste workflow between AI and platforms
- Different formats required for each social network
- Manual image generation and attachment
- Knowledge base searches separate from AI conversations
What I built
Integrated Publishing
MCP servers expose platform capabilities directly to Claude. Discuss content, draft it, and publish, all in one conversation. No context switching, no copy-paste, no friction.
- Instagram MCP: Images, videos, reels, carousels
- LinkedIn MCP: Articles and posts
- X/Twitter MCP: Tweets and threads
- Obsidian Search: Semantic knowledge base queries
How it fits together
Architecture
TypeScript MCP servers implement OAuth flows and API integrations. Secure credential storage in macOS Keychain. Rate limiting and error handling built in.
Timeline
How it went
- Sep 2024Instagram MCP first version
- Oct 2024LinkedIn and X servers added
- Nov 2024Obsidian semantic search MCP
- Dec 2024Gemini image generation MCP
- Jan 2025Open source release
What it taught me
Lessons
- Reducing friction transforms how often you share insights
- MCP is the right abstraction for extending AI capabilities
- OAuth flows are the hardest part: once solved, integration is easy
- Shared tools accelerate the entire MCP ecosystem
Stack
Connected work