I'm building Aiko, an AI tools marketplace that solves the headaches of managing MCP (Model Context Protocol) configurations and credentials.

After working with and building many MCP servers, I kept running into the same frustrating issues:

- Managing multiple environments (production vs development) meant either duplicate server configs or constantly updating API keys

- Too many enabled tools either hit context limits or caused LLMs to hallucinate when choosing the right tool

- MCP servers often expose dozens of tools, but I only wanted a subset and had to configure this separately for Claude, Cursor, and every other AI assistant

- Most MCP aggregators are third-party hosted, which meant sharing sensitive credentials with external services

Aiko addresses these problems by providing a centralized way to manage your MCP configurations, credentials, and tool selections. You can easily switch between environments (profile management), selectively enable/disable tools globally, and all your credentials are encrypted (with MacOS Keychain) and stored locally.

I've also added some interesting MCP servers as part of this marketplace, including one that converts any OpenAPI v3 spec into an MCP server with full authentication support ().

The platform is still in beta, and I'm actively working on expanding the tool marketplace. Some features like Gmail and Calendar integrations are currently under security assessment but you can see them in action in the demo video.

Would love to get feedback from the HN community, especially from others who've been wrestling with MCP management!

- Demo (Google Calendar + Gmail + Memory working together): https://www.youtube.com/watch?v=ZgEy6Y1kfn4

- Website: https://getaiko.app

- OpenAPI to MCP - weather example: https://x.com/getaikoapp/status/1945278307496235482