arazzo-mcp-gen
arazzo-mcp-gen is a CLI tool that turns an Arazzo specification and its referenced OpenAPI files into a fully Dockerized Python MCP (Model Context Protocol) server. Each Arazzo workflow becomes an MCP tool that any AI agent can call.
에이전트 강화
게시자가 제공하는 설치 방식입니다. PlugMCP 엔드포인트가 아닙니다.
원격 URL인가요, 내 컴퓨터에서 실행하는 명령인가요? 두 설치 방식의 차이
PlugMCP에서 실행
A new server you create. It is not this listing's official endpoint. Use the publisher install when you want their URL and their credentials.
실행게시자 제공
arazzo-mcp-gen is a CLI tool that turns an Arazzo specification and its referenced OpenAPI files into a fully Dockerized Python MCP (Model Context Protocol) server. Each Arazzo workflow becomes an MCP tool that any AI agent can call.
Automatically generate fully functional Model Context Protocol (MCP) servers from Arazzo specifications. Features built-in utilities for workflow validation, inspection, and visualization.
Given a folder containing:
.yaml file (describes multi-step API workflows).yaml files (describe individual API operations)…the CLI will:
Download the archive for your platform from the Releases page, or use the quick install commands below. All platforms ship as .zip.
# For macOS (Apple Silicon)
curl -L https://github.com/wso2/arazzo-mcp-generator/releases/download/0.1.0/arazzo-mcp-gen-0.1.0-darwin-arm64.zip -o arazzo-mcp-gen.zip
unzip arazzo-mcp-gen.zip
sudo mv arazzo-mcp-gen-darwin-arm64 /usr/local/bin/arazzo-mcp-gen
rm arazzo-mcp-gen.zip
# For macOS (Intel)
curl -L https://github.com/wso2/arazzo-mcp-generator/releases/download/0.1.0/arazzo-mcp-gen-0.1.0-darwin-amd64.zip -o arazzo-mcp-gen.zip
unzip arazzo-mcp-gen.zip
sudo mv arazzo-mcp-gen-darwin-amd64 /usr/local/bin/arazzo-mcp-gen
rm arazzo-mcp-gen.zip
# For Linux (x86_64)
curl -L https://github.com/wso2/arazzo-mcp-generator/releases/download/0.1.0/arazzo-mcp-gen-0.1.0-linux-amd64.zip -o arazzo-mcp-gen.zip
unzip arazzo-mcp-gen.zip
sudo mv arazzo-mcp-gen-linux-amd64 /usr/local/bin/arazzo-mcp-gen
rm arazzo-mcp-gen.zip
# For Linux (ARM64)
curl -L https://github.com/wso2/arazzo-mcp-generator/releases/download/0.1.0/arazzo-mcp-gen-0.1.0-linux-arm64.zip -o arazzo-mcp-gen.zip
unzip arazzo-mcp-gen.zip
sudo mv arazzo-mcp-gen-linux-arm64 /usr/local/bin/arazzo-mcp-gen
rm arazzo-mcp-gen.zip
Each archive contains the binary, LICENSE, and README.md. Binaries are statically linked and have no runtime dependencies.
macOS note: the binaries are unsigned and un-notarized. If Gatekeeper blocks the binary, clear the quarantine flag:
xattr -d com.apple.quarantine /usr/local/bin/arazzo-mcp-gen
# Download and extract (PowerShell)
Invoke-WebRequest -Uri https://github.com/wso2/arazzo-mcp-generator/releases/download/0.1.0/arazzo-mcp-gen-0.1.0-windows-amd64.zip -OutFile arazzo-mcp-gen.zip
Expand-Archive -Path arazzo-mcp-gen.zip -DestinationPath .
# Move arazzo-mcp-gen.exe to a directory in your PATH, or run it directly
arazzo-mcp-gen --version
# arazzo-mcp-gen version v0.1.0
To verify an archive against the release's checksums.txt:
shasum -a 256 -c checksums.txt --ignore-missing
validateValidates an Arazzo specification for correctness and completeness.
Uses Spectral (via npx @stoplight/spectral-cli) with the official spectral:arazzo ruleset as the primary validator when available. Falls back to the built-in Go validator when Node.js is not installed, showing install instructions.
arazzo-mcp-gen validate -f <file-or-folder>
Examples
# Validate a folder (auto-detects the Arazzo file)
arazzo-mcp-gen validate -f ./my-arazzo-folder
# Validate a single file
arazzo-mcp-gen validate -f ./workflow.yaml
# Validate and also check that remote OpenAPI URLs are reachable
arazzo-mcp-gen validate -f ./my-arazzo-folder --check-remote
# Strict mode: fail if there are any warnings
arazzo-mcp-gen validate -f ./my-arazzo-folder --strict
What it checks (Spectral ruleset)
workflowId and stepId valuesoperationId, operationPath, workflowId) are present and validname, in, and value fieldsonSuccess / onFailure action namesdependsOn cross-referencesAdditional built-in checks (always run)
--check-remote)$statusCode criteria that are AND-ed together (a common mistake)Exit codes
inspectParses and prints a detailed, colour-coded overview of an Arazzo spec — without generating anything. Use this to understand a spec or debug step-flow routing before generating an MCP server.
arazzo-mcp-gen inspect -f <file-or-folder>
Examples
# Inspect a folder (auto-detects the Arazzo file)
arazzo-mcp-gen inspect -f ./my-arazzo-folder
# Inspect a specific file
arazzo-mcp-gen inspect -f ./workflow.yaml
Output includes
onSuccess / onFailure routing with conditions (GOTO, END, RETRY)visualizeGenerates a Mermaid flowchart diagram of the Arazzo spec's workflow logic. By default opens the rendered diagram in your browser (no extra tools needed). Can also save to a file.
arazzo-mcp-gen visualize -f <file-or-folder> [-o <output-file>]
Alias: viz
Examples
# Open diagram in browser (default)
arazzo-mcp-gen visualize -f ./my-arazzo-folder
# Save to GitHub-renderable Markdown
arazzo-mcp-gen visualize -f ./workflow.yaml -o diagram.md
# Save raw Mermaid source
arazzo-mcp-gen visualize -f ./my-arazzo-folder -o flow.mmd
# Short alias
arazzo-mcp-gen viz -f ./my-arazzo-folder
Diagram shows
onSuccess / onFailure branches labelled with conditionsgoto referencesPaste any
.mmdfile into mermaid.live for a shareable interactive link.
mcp-server generateThe main command. Reads your Arazzo + OpenAPI files, generates a Python MCP server, and builds a Docker image.
arazzo-mcp-gen mcp-server generate -f <file-or-folder> [flags]
Examples
# From a folder (auto-detects the Arazzo file)
arazzo-mcp-gen mcp-server generate -f ./my-arazzo-folder
# From a single Arazzo file directly
arazzo-mcp-gen mcp-server generate -f ./my-arazzo-folder/workflow.arazzo.yaml
# Custom port
arazzo-mcp-gen mcp-server generate -f ./my-arazzo-folder -p 8080
# Inspect generated files after build
arazzo-mcp-gen mcp-server generate -f ./workflow.arazzo.yaml -p 8080 -o ./artifacts
Input requirements
-f is a folder: it must contain exactly one .yaml/.yml file with a top-level arazzo: key-f is a file: point directly to the Arazzo file; the folder can contain multiple Arazzo filessourceDescriptions[].url must be in the same folder as the Arazzo fileinfo.title, info.version, and at least one workflowWhat it does
mcp_server.py — each workflow becomes a @mcp.tool() function with typed parametersDockerfile using python:3.11-slimdocker build to produce a tagged imagedocker run command to start the serverRunning the generated server
docker run -p 5000:5000 <image-name>
If your workflow uses HTTPS endpoints with self-signed or otherwise invalid TLS certificates, run the image with the following environment variable to disable certificate verification:
docker run -e ARAZZO_DISABLE_TLS_VERIFY=1 -p 5000:5000 <image-name>
The MCP endpoint is available at http://localhost:5000/mcp.
Scenario: You have an OpenAPI spec for a pet store API and want to expose a "check if a pet exists, then create or update it" workflow as an MCP tool for an AI agent.
Create a folder containing your Arazzo specification and its referenced OpenAPI files:
pet-project.petstore_workflow.yaml) inside it..yaml files referenced in the Arazzo spec are also in this folder.pet-project/
├── petstore_workflow.yaml ← Your Arazzo spec
└── petstore_openapi.yaml ← Your OpenAPI spec
arazzo-mcp-gen validate -f .
Expected output (Spectral available):
Validating: /path/to/pet-project/petstore_workflow.yaml
────────────────────────────────────────────────────────────
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
Validation Result: PASSED
✓ All arazzo rules passed
─ Validated using Spectral (spectral:arazzo ruleset)
Fix any errors reported before continuing. Warnings are informational; use --strict to treat them as errors in CI.
arazzo-mcp-gen inspect -f .
Review the printed summary to confirm:
operationId that matches your OpenAPI specarazzo-mcp-gen visualize -f .
Your browser opens with an interactive Mermaid flowchart. Check the branching logic visually — this is especially useful for multi-step workflows with onSuccess / onFailure routing.
To save it:
# As a Markdown file (renders on GitHub)
arazzo-mcp-gen visualize -f . -o flow.md
Make sure Docker is running, then:
arazzo-mcp-gen mcp-server generate -f . -p 5000 -o ./artifacts
Expected output:
Validating input folder...
Found Arazzo spec: Pet Upsert Workflow (V3) with 1 workflow(s)
Generating MCP server code...
Building Docker image...
[+] Building 12.3s (10/10) FINISHED
╔════════════════════════════════════════════════════════════════════════╗
║ ✅ MCP Server image built successfully! ║
║ ║
║ Image: pet-upsert-workflow-v3-mcp-server ║
║ Run: docker run -p 5000:5000 pet-upsert-workflow-v3-mcp-server ║
║ URL: http://localhost:5000 ║
║ ║
║ If TLS verification must be disabled for self-signed HTTPS endpoints, ║
║ run the image with: -e ARAZZO_DISABLE_TLS_VERIFY=1 ║
║ ║
║ Build artifacts saved to: /path/to/pet-project/artifacts ║
╚════════════════════════════════════════════════════════════════════════╝
Copy the docker run command from the output and run it:
docker run -p 5000:5000 pet-upsert-workflow-v3-mcp-server
The server is now live at http://localhost:5000/mcp in stateless HTTP mode. To connect it to an MCP client like Claude Desktop, you can use supergateway to bridge the HTTP endpoint:
{
"mcpServers": {
"my-mcp-server": {
"command": "npx",
"args": [
"-y",
"supergateway",
"--streamableHttp",
"http://localhost:5000/mcp"
]
}
}
}
Note: Replace
http://localhost:5000/mcpwith the endpoint shown in your terminal if you used a different port.
The AI agent can now call your Arazzo workflows as tools. The tool executes the full multi-step logic internally and returns the final result.
Inspect with --output / -o ./artifacts:
artifacts/
├── mcp_server.py ← FastMCP server; each workflow = @mcp.tool()
├── Dockerfile ← python:3.11-slim image; EXPOSEs your port
└── arazzo/
├── petstore_workflow.yaml ← copy of your Arazzo spec
└── openapi.yaml ← copy of referenced OpenAPI spec(s)
To get started, create a folder for your project and save the following as petstore_workflow.yaml inside it. This is a ready-to-use Arazzo spec targeting the public Petstore v3 API — it checks whether a pet exists by ID, updates its name if found, or creates it if not.
arazzo: "1.0.0"
info:
title: Pet Upsert Workflow (V3)
summary: A sample workflow that conditionally creates or updates a pet using Petstore V3
description: Workflow targeting Petstore V3 API. Takes an id and name - renames the pet if it exists, creates it if not.
version: 1.0.0
sourceDescriptions:
- name: petstoreApiV3
url: https://petstore3.swagger.io/api/v3/openapi.json
type: openapi
workflows:
- workflowId: ensurePetExistsV3
summary: Check if a pet exists by ID; update its name if found, create it if not.
description: This workflow demonstrates conditional logic based on API responses. It first checks if a pet with the given ID exists. If it does, it updates the pet's name. If it doesn't, it creates a new pet with the provided ID and name.
inputs:
type: object
properties:
petId: { type: integer }
newName: { type: string }
steps:
- stepId: checkStep
description: Check if the pet exists and route accordingly.
operationId: getPetById
parameters:
- name: petId
in: path
value: $inputs.petId
successCriteria:
- condition: $statusCode == 200
# Branch based on which status code was returned
onSuccess:
- name: petFoundRouteToUpdate
criteria:
- condition: $statusCode == 200
type: goto
stepId: updateStep
# Retry on true server errors
onFailure:
- name: retryOnServerError
criteria:
- condition: $statusCode >= 500
type: retry
retryAfter: 5
- stepId: createStep
description: Pet not found - create it with the given id and name.
operationId: addPet
requestBody:
contentType: application/json
payload:
id: $inputs.petId
name: $inputs.newName
category:
id: 1
name: Dogs
photoUrls:
- "https://example.com/pet.jpg"
tags:
- id: 0
name: string
status: "available"
onSuccess:
- name: endAfterCreation
type: end
- stepId: updateStep
description: Pet found - rename it using a full PUT update.
operationId: updatePet
requestBody:
contentType: application/json
payload:
id: $inputs.petId
name: $inputs.newName
category:
id: 1
name: Dogs
photoUrls:
- "https://example.com/pet.jpg"
tags:
- id: 0
name: string
status: "available"
onSuccess:
- name: endAfterUpdate
type: end
Once you have this file saved, follow the End-to-End Walkthrough to validate, inspect, and generate an MCP server from it.
Copyright (c) 2026, WSO2 LLC. (https://www.wso2.com).
Automatically generate fully functional Model Context Protocol (MCP) servers from Arazzo specifications. Features built-in utilities for workflow validation, inspection, and visualization.
도구
아직 도구 목록이 없습니다. 원격 서버는 첫 상태 확인 후에 도구 목록이 생깁니다.
FAQ
이 페이지에서 Cursor, Claude Desktop, Claude Code 또는 VS Code용 stdio 설정을 복사합니다. PlugMCP 프록시가 아니라 게시자의 설치 방식을 사용합니다.
이 레코드에는 필수 환경 변수가 등록되어 있지 않습니다.
Arazzo MCP Generator은(는) stdio를 통해 로컬에서 실행됩니다.
마지막 확인: 2026-10-28. Automatically generate fully functional Model Context Protocol (MCP) servers from Arazzo specifications. Features built-in utilities for workflow validation, inspection, and visualization.