Skip to main content
Glama
Work90210

APIFold

by Work90210
OverviewSchemaRelated ServersScoreDiscussions
TypeScript
Hybrid

CI License Transformer npm Node

APIFold reads an OpenAPI 3.x or Swagger 2.x specification and generates a live, production-ready MCP server endpoint. AI agents — Claude, Cursor, Copilot, or any MCP-compatible client — can connect immediately. Tool calls execute real HTTP requests to real upstream APIs with securely stored credentials. No stubs, no mocks, no glue code.

Quick Start

CLI (fastest)

npx apifold serve ./your-openapi-spec.yaml --base-url https://api.example.com

One command. Your spec becomes a running MCP server with SSE transport on localhost:3000. Connect Claude or Cursor immediately.

From the API Registry

npx apifold serve --registry stripe --base-url https://api.stripe.com

Ships with 8 pre-configured API specs: Stripe, GitHub, Slack, HubSpot, Twilio, OpenAI, Notion, and Petstore.

Hosted Platform

  1. Sign up at apifold.dev

  2. Import a spec (URL, file upload, or browse the registry)

  3. Copy the connection snippet into Claude Desktop or Cursor

  4. Your MCP server is live with a unique endpoint URL

Related MCP server: api-to-mcp

Features

  • OpenAPI 3.x + Swagger 2.0 — auto-converts Swagger 2.0 specs transparently

  • CLI tool — npx apifold serve for zero-config local MCP servers

  • API registry — one-click deploy from 8 curated API specs

  • OAuth 2.0 — Authorization Code (PKCE), Client Credentials, 8 provider presets, auto token refresh

  • Access profiles — tool-level permissions (Read Only / Read-Write / Full Access)

  • Analytics — call volume, latency percentiles, error breakdown, usage quotas

  • Custom domains — use mcp.yourcompany.com with DNS verification

  • Unique endpoint IDs — cryptographic, unguessable URLs per server

  • SSE + Streamable HTTP — both MCP transport modes supported

  • Vault encryption — AES-256-GCM for all stored credentials

  • SSRF protection — DNS pinning, redirect blocking, private IP rejection

CLI

npm install -g @apifold/cli

Command

Description

apifold serve <spec>

Start an MCP server from an OpenAPI spec

apifold serve --registry stripe

Start from a registry spec

apifold transform <spec>

Output MCP tool definitions as JSON

apifold validate <spec>

Parse-only validation with warnings

apifold init [spec]

Generate an apifold.config.yaml template

Config file

# apifold.config.yaml
spec: ./openapi/stripe.yaml
port: 3001
transport: sse
baseUrl: https://api.stripe.com
auth:
  type: bearer
  token: ${STRIPE_API_KEY}
filters:
  tags: [payments, customers]
  methods: [get, post]

Self-Hosting

Development

git clone https://github.com/Work90210/APIFold.git
cd APIFold
pnpm install
cp .env.example .env
docker compose -f infra/docker-compose.dev.yml up -d
pnpm dev

Open http://localhost:3000 to access the dashboard.

Production

cp .env.example .env
# Edit .env with your production values
docker compose -f infra/docker-compose.yml up -d

Using the Transformer Library

The core conversion logic is a standalone MIT-licensed npm package:

npm install @apifold/transformer
import { parseSpec, transformSpec, autoConvert } from "@apifold/transformer";

// Auto-convert Swagger 2.0 if needed
const { spec } = await autoConvert(rawSpec);

// Parse and transform
const parsed = parseSpec({ spec });
const { tools } = transformSpec({ spec: parsed.spec });
// tools = MCPToolDefinition[] ready for any MCP server

Architecture

graph LR
    Spec["OpenAPI / Swagger\nSpec"] --> Transformer

    subgraph MT["APIFold"]
        Transformer["Transformer\n(MIT lib)"]
        Transformer --> Runtime["Runtime\nExpress + SSE"]
        Web["Web App\nNext.js"] --> DB["Postgres 16\nRedis 7"]
        Runtime --> DB
        CLI["CLI Tool\nnpx apifold"]
        Registry["API Registry\n8 specs"]
    end

    Runtime --> Upstream["Upstream API\n(Stripe, GitHub, etc.)"]
    Agent["AI Agent\nClaude, Cursor,\nCopilot, etc."] -- "MCP / SSE" --> Runtime

Component

Path

Description

Transformer

packages/transformer

OpenAPI to MCP conversion. Pure functions. MIT licensed.

Runtime

apps/runtime

Express MCP server with SSE/HTTP, OAuth token refresh, access profiles.

Web App

apps/web

Next.js dashboard. Import specs, manage credentials, analytics, custom domains.

CLI

apps/cli

Standalone CLI tool. npx apifold serve for local MCP servers.

Registry

packages/registry

Curated catalog of 8 validated API specs with one-click deploy.

Types

packages/types

Shared TypeScript type definitions.

UI

packages/ui

Design system and component library.

Available Commands

Command

Description

pnpm dev

Start all services with hot-reload

pnpm build

Build all packages

pnpm test

Run tests

pnpm lint

Lint all packages

pnpm typecheck

Type-check all packages

pnpm format

Format all files with Prettier

pnpm db:migrate

Run database migrations

pnpm db:studio

Open Drizzle Studio

Contributing

All code contributions must go through a pull request and be approved before merging. See the contribution guide.

Want to add an API to the registry? See packages/registry/CONTRIBUTING.md.

Security

For security issues, please refer to our security policy. Do not post security vulnerabilities as public GitHub issues.

License

Available Tools

2 tools
get_server_endpointB

Get the SSE endpoint URL for a specific APIFold MCP server by slug.

ParametersJSON Schema
NameRequiredDescriptionDefault
slugYesThe server slug (e.g. github, stripe, slack, openai, notion)
userKeyYesYour APIFold user key

TDQS

B3.3/5.0
Behavior2/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

No annotations are provided, so the description must carry the full burden of behavioral disclosure. It mentions what is retrieved (SSE endpoint URL) but omits error handling (e.g., invalid slug), side effects, caching behavior, or whether this operation incurs costs or rate limits.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness5/5

Is the description appropriately sized, front-loaded, and free of redundancy?

The description is a single, efficient sentence that front-loads the action verb. There is no redundant or wasted text; every word contributes to understanding the tool's function.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness3/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

While adequate for a simple two-parameter lookup tool, the description lacks information about the return value format (critical given no output schema exists) and error conditions. It meets minimum viability but leaves operational gaps.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters3/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

With 100% schema description coverage, the schema already comprehensively documents both parameters (including examples for 'slug'). The description adds no additional parameter semantics beyond implying the 'slug' parameter via 'by slug', so it meets the baseline expectation.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose5/5

Does the description clearly state what the tool does and how it differs from similar tools?

The description uses a specific verb ('Get') and resource ('SSE endpoint URL') and clearly scopes the operation to a specific server identified 'by slug'. This effectively distinguishes it from sibling tool 'list_servers' (which presumably enumerates servers rather than retrieving a specific endpoint).

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines2/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

The description provides no guidance on when to use this tool versus alternatives. It fails to mention that 'list_servers' should likely be used first to discover valid slugs, or under what circumstances an agent should call this endpoint.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

list_serversA

List all 18 free hosted APIFold MCP servers (GitHub, Stripe, Slack, OpenAI, Notion, and more). Each server proxies a REST API — bring your own API key.

ParametersJSON Schema
NameRequiredDescriptionDefault

No parameters

TDQS

A4.2/5.0
Behavior4/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

With no annotations provided, the description carries the full burden of behavioral disclosure. It successfully reveals critical ecosystem constraints: the servers are 'free hosted,' 'proxy a REST API,' and require users to 'bring your own API key.' However, it omits details about response format, caching, or rate limiting.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness5/5

Is the description appropriately sized, front-loaded, and free of redundancy?

The description consists of two highly efficient sentences. The first front-loads the core deliverable (the 18 specific servers), while the second adds essential context about the REST proxy architecture and API key requirements. No words are wasted.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness4/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

Given the low complexity (zero parameters, no nested objects) and lack of output schema, the description adequately explains what the tool returns (inventory of 18 servers) and their operational model. It could be improved by hinting at the return structure (e.g., array of server objects), but it is sufficiently complete for agent selection.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters4/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

Per the evaluation rules, tools with zero parameters receive a baseline score of 4. The input schema is an empty object with no parameters to describe, and the description appropriately focuses on the return value rather than non-existent inputs.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose5/5

Does the description clearly state what the tool does and how it differs from similar tools?

The description clearly states the tool 'List[s] all 18 free hosted APIFold MCP servers' with specific examples (GitHub, Stripe, Slack, etc.). The plural 'servers' and enumeration of all 18 naturally distinguishes it from sibling 'get_server_endpoint' (singular), making the scope unambiguous.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines3/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

The description implies this is a discovery/initialization tool ('List all'), but provides no explicit guidance on when to use this versus 'get_server_endpoint' or prerequisites for using the listed servers. The usage context is inferred from the verb 'List' but not stated explicitly.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

Install Server
RayrunBuild, deploy, and operate MCP servers without becoming an infrastructure team.ray.runAd

TDQS

A3.8/5.0
Disambiguation5/5

The two tools have clearly distinct purposes: one lists all available servers (discovery), while the other retrieves a specific connection endpoint (retrieval). No functional overlap exists between listing a catalog and fetching a specific URL.

Naming Consistency5/5

Both tools follow a consistent verb_noun pattern using snake_case (get_server_endpoint, list_servers). The verb choices ('get' vs 'list') accurately reflect the read operations and maintain predictable conventions.

Tool Count3/5

With only 2 tools, the surface feels minimal/thin despite the server connecting to 18 different API proxies. While functional for pure discovery (list + get), the count is borderline for the implied richness of the domain.

Completeness4/5

Covers the core read-only lifecycle for a directory service: discovery (list) and connection (get endpoint). Minor gap in filtering/search capabilities for the 18 servers, though listing all is acceptable at this scale.

Maintenance

ActivityInactive
ResponsivenessSyncing

Resources

Unclaimed servers have limited discoverability.

Looking for Admin?

If you are the server author, to access and configure the admin panel.

Related MCP Connectors

Related MCP Servers

  • mcpify

    qualuo
    A
    license
    Not graded
    quality
    C
    maintenance
    Turns any OpenAPI specification into a fully working MCP server with a single command, enabling AI agents to call APIs without writing any glue code.
    13
    MIT

  • api-to-mcp

    sgaluza
    A
    license
    Not graded
    quality
    D
    maintenance
    Turns any OpenAPI REST or GraphQL API into an MCP server, allowing AI assistants like Claude to interact with APIs without writing code.
    165
    MIT

  • akyn-ai

    IlyesTal
    A
    license
    A
    quality
    D
    maintenance
    Turn any data source into an MCP server in 5 minutes. Build knowledge bases that AI assistants like Claude and Cursor can query directly.
    2
    28
    23
    MIT

Appeared in Searches

Latest Blog Posts

MCP directory API

We provide all the information about MCP servers via our MCP API.

curl -X GET 'https://glama.ai/api/mcp/v1/servers/Work90210/APIFold'

If you have feedback or need assistance with the MCP directory API, please join our Discord server