APICHAP, based in Linz, Austria, offers a service for building and hosting “MCP servers” ─ the layer that connects AI to your APIs and databases ─ without writing code. MCP (Model Context Protocol) here refers to the connection standard that lets multiple AI agents such as Claude, ChatGPT, and Copilot access external tools and data in a unified way. Note that MCP-Builder.ai is a tool for creating MCP servers, not a company that publishes an official MCP server of its own. Upload an OpenAPI specification and it auto-generates tool definitions, turning REST/GraphQL APIs, SQL/NoSQL databases, and CSV/Excel/XML files into something an AI can operate in a matter of minutes. A Product Hunt listing dates the launch to around June 17, 2026, though the official site states no clear launch date, so confidence here is moderate.
Key Features
- No-code MCP server generation: Upload an OpenAPI specification and tool definitions are generated automatically, standing up an MCP server in seconds
- Broad data source support: REST/GraphQL/SOAP/JSON-RPC/XML-RPC APIs, databases including PostgreSQL/MySQL/MariaDB/MSSQL/Oracle/SQLite, and CSV/Excel/XML files
- Multiple transport protocols: STDIO and SSE (Server-Sent Events) are supported. HTTP-Streamable transport is still in development at the time of writing
- Secure authentication: API key authentication and OAuth 2.0 are supported, with connections to sensitive data sources in mind
- Multiple deployment options: Run locally over STDIO, deploy on-premises as a container, or use cloud hosting on Azure Container Instance (an add-on)
- Marketplace: A mechanism for partners to publish and share the MCP servers they build
Pricing
Pricing is a monthly subscription, with higher tiers unlocking more tools and additional protocols.
| Plan | Monthly | What you get |
|---|---|---|
| Starter | $0 | 1 MCP server, 1 tool, STDIO only. Aimed at public community servers |
| Pro | $25 | 1 MCP server, 6 tools, STDIO/SSE. Aimed at private production use |
| Scale | $60 | 1 MCP server, unlimited tools, STDIO/SSE, OAuth 2.0 |
| Cloud Hosting Add-on | $55 | Cloud hosting on Azure Container Instance (1 vCPU / 4 GB RAM). Optional add-on |
Pricing reflects information as of September 2026. Check the official site for current rates.
Pros and Cons
✅ Pros
- Generates tool definitions in seconds from documentation you already have, such as an OpenAPI spec
- Wide range of supported data sources, from REST/GraphQL to databases and file formats
- Implements the main MCP specifications: STDIO, SSE, and OAuth 2.0
- A free Starter plan makes small-scale experiments easy to begin
⚠️ Cons
- The free plan is limited to 1 tool and STDIO only, so real use requires Pro or above
- Cloud hosting is a separate $55/month add-on, making production costs harder to predict
- HTTP-Streamable transport is not yet supported (in development at the time of writing)
- A young service ─ Japanese-language coverage and case studies remain scarce
How It Compares
| Criteria | MCP-Builder.ai | MCPCore | Zapier MCP | Official SDKs (Python/TypeScript) |
|---|---|---|---|---|
| Provider | APICHAP | ─ | Zapier | Anthropic and others |
| Primary use | Auto-convert APIs/DBs into MCP servers | Same (cloud-based no-code building) | Add an MCP endpoint to an existing automation platform | Implement an MCP server in code |
| How you build | Auto-generated from an OpenAPI upload | No-code | Configured within Zapier | By hand (coding required) |
| Supported data sources | Wide: APIs, DBs, file formats | ─ | Mostly Zapier integrations | Anything (depends on your implementation) |
| Best suited for | A broad audience, including non-engineers | No-code-oriented teams | Teams already using Zapier | Engineers and teams that prioritize customization |
Who It’s For
- Anyone responsible for making an internal REST API safely callable from AI agents without writing code
- People who want to expose an existing PostgreSQL or MySQL database as a tool for an LLM
- Developers who need to quickly turn a third-party SaaS with nothing but an OpenAPI spec into an MCP server and connect it to Claude, ChatGPT, or Copilot
- Builders who want to distribute their own MCP servers through a marketplace and are interested in monetizing them
Summary
MCP-Builder.ai is not a service that gives you an MCP server ─ it is a building and hosting platform for producing “MCP servers an AI can use” from an OpenAPI spec or database connection details, without code. Its strengths are the breadth of supported data sources and coverage of STDIO/SSE/OAuth 2.0. But the free plan is tightly constrained at 1 tool and STDIO only, production use requires a paid Pro plan or higher, and always-on cloud operation needs yet another paid add-on. Start by turning one of your own APIs into an MCP server on the free tier, confirm an AI agent can actually call it, then pick a plan based on the features you find you need.