N8n Unified Mcp Server

devops-infra MCP Server

Production-grade Unified MCP Server architected to bridge enterprise n8n automation with agentic LLM ecosystems. Engineered for high-scale reliability with structured logging, multi-transport support (SSE/STDIO), and context-aware orchestration. Accelerating AI-native business infrastructure via secure, deterministic, and scalable workflow control.

Verified
devops-infradevops-infra
5 views0 stars0 forks

Why This Matters

Discovered via github-topic:claude-mcp and last synced 3mo ago.

Verified
Source
github-topic:claude-mcp
Stars
0
Last synced
3mo ago
Install
Check source

Install

Install instructions not detected yet

Check the source repository for the latest setup steps.

View source instructions
180
Tools
0
Resources
0
Prompts
Standard I/O
Transport

Available Tools (180)

AC-05

`validate_workflow` catches structural errors before deploy

Python

3.12

Disk

200 MB

13

[🔍 Validation Summary](#-validation-summary)

Section

#

Criterion

Verification Method

AC-09

n8n connectivity failure returns `{"status": "degraded"}`

AC-01

Server starts on port 8000 and responds at `/health`

AC-10

Docker image builds and runs without errors

fastapi

ASGI web framework

T-01

Server Startup

T-10

Validate Workflow — Invalid

T-19

Retry Logic

512mb

Comfortable for Python + httpx pool

M-08

Bearer auth middleware

M-17

OpenTelemetry tracing

T-15

Docker Build

T-08

Cache — Second Request

T-17

Stdio Mode

min_machines_running

`1`

M-15

Redis cache backend

12

[🧪 Testing & Validation](#-testing--validation)

3

Ensure zero data loss during workflow operations

7

Provide Prometheus-compatible observability

14

[🧰 Verification Testing Tools & Commands](#-verification-testing-tools--commands)

5

Achieve sub-100 ms response for cached operations

AC-03

Bearer token authentication rejects invalid tokens

Package

Purpose

passlib

Password / secret hashing utilities

MCP_PORT

No

CACHE_TTL_TEMPLATES

No

ENABLE_AUTO_FIX

No

GET

`/redoc`

Created

Workflow creation via n8n API

n8n_reachable

`true` / `false`

delete_workflow

`workflow_id`

get_execution_details

`execution_id`

list_node_categories

validate_connection

`source_node`, `target_node`

FastMCP

Latest

Docker

24.x

T-06

List Workflows

T-13

Metrics Endpoint

T-20

n8n Degraded

shared-cpu-1x

Free tier; sufficient for MCP workload

Quarterly

Regenerate in n8n UI, update `.env`

M-03

All 5 execution & debugging tools

M-10

Structured logging + Prometheus metrics

8

Allow optional Redis cache for horizontal scaling

15

[🧯 Troubleshooting & Debugging](#-troubleshooting--debugging)

6

Support cloud-native deployment with zero cold starts

AC-04

`/health` and `/docs` endpoints bypass authentication

AC-11

Fly.io deployment is reachable at HTTPS endpoint

mcp

Model Context Protocol core runtime

typing-extensions

Backported type hints

LOG_LEVEL

No

REDIS_URL

No

ENABLE_CONTEXT_MEMORY

No

Header

Description

Deleted

Workflow or execution deleted

status

`"healthy"` / `"degraded"`

activate_workflow

`workflow_id`

analyze_workflow_errors

`workflow_id`

get_node_example

`node_type`

list_credentials

Latest

Bearer token validation

ID

Area

T-07

Cache — First Request

T-14

Cache Clear

Symptom

Likely Cause

M-04

All 5 node documentation tools (40+ nodes)

M-11

Multi-stage Docker build + Compose

Category

Requirement

10

[🧱 Tech Stack & Architecture](#-tech-stack--architecture)

1

Provide a single MCP endpoint for all AI clients

AC-06

`auto_fix_workflow` repairs missing IDs and positions

Component

Minimum

httpx

Latest

N8N_API_KEY

✅ Yes

CACHE_ENABLED

No

N8N_API_TIMEOUT

No

RATE_LIMIT_ENABLED

No

POST

`/cache/clear`

X-Response-Time

Server processing time in milliseconds

Degraded

n8n connectivity lost; health reports `degraded`

list_workflows

`active_only`, `tag`, `limit`

execute_workflow

`workflow_id`, `data`

get_latest_execution

`workflow_id`

get_next_step_recommendations

get_credential_types_reference

v2

Type-safe env var management

T-02

Health Endpoint

T-09

Validate Workflow — Valid

T-16

Docker Run

Setting

Value

TTL

Rationale

M-06

AI intelligence tools + session context (5 tools)

M-13

Dual transport (HTTP/SSE + stdio)

11

[🛠️ Workflow & Implementation](#️-workflow--implementation)

AC-07

Cache returns hit on second identical request within TTL

tenacity

Retry logic with exponential backoff

N8N_BASE_URL

✅ Yes

CACHE_TTL_WORKFLOWS

No

N8N_API_RETRIES

No

RATE_LIMIT_REQUESTS

No

X-Server

Server name identifier

Value

Meaning

get_workflow

`workflow_id`

duplicate_workflow

`workflow_id`, `new_name`

validate_workflow

`workflow_json`

analyze_workflow

`workflow_id`

Layer

Technology

Tenacity

Latest

T-03

Auth — No Token

primary_region

`sin` (Singapore)

Scenario

Recommendation

Milestone

Description

M-07

Credentials tools (3 tools)

M-14

Client config docs (5 clients)

2

Expose all n8n REST API v1 capabilities as MCP tools

9

[🚀 Features](#-features)

Objective

Priority

AC-12

Prometheus metrics are exposed in correct text format

Variable

Required

LOG_FORMAT

No

WORKERS

No

ENABLE_TEMPLATE_SEARCH

No

X-Request-ID

Unique request identifier for tracing

Duplicate

Creating a workflow with a name that already exists

Tool

Key Parameters

deactivate_workflow

`workflow_id`

get_execution_metrics

`workflow_id`

get_expression_guide

`topic`

get_session_context

get_credential_schema

`credential_type`

Rule

Description

hard_limit

`200` connections

Weekly

Large log files indicate unexpected traffic

M-05

Validation + auto-fix pipeline (4 tools)

M-12

Fly.io deployment config

4

Enable context-aware AI assistance across tool calls

AC-02

All 30 MCP tools are registered and discoverable

python-jose

JWT / bearer token security

MCP_HOST

No

CACHE_TTL_EXECUTIONS

No

ENABLE_WORKFLOW_VALIDATION

No

Method

Endpoint

Success

All successful GET/POST responses

success

`true` / `false`

update_workflow

`workflow_id`, `workflow_json`

list_executions

`workflow_id`, `status`, `limit`

get_node_documentation

`node_type`

build_workflow_template

`description`

search_workflow_templates

`query`

FastAPI

0.115+

T-05

Auth — Bypass Health

T-12

Session Context

Monthly

`pip install -r requirements.txt --upgrade`

M-02

All 9 workflow management tools

M-09

TTL cache with background eviction

M-16

JWT auth with expiry

AC-08

Session context persists across multiple tool calls

RAM

256 MB

structlog

Latest

MCP_BEARER_TOKEN

⚠️ Strongly recommended

CACHE_TTL_NODES

No

N8N_MAX_CONNECTIONS

No

RATE_LIMIT_WINDOW

No

Code

Meaning

valid

`true` / `false`

create_workflow

`workflow_json`

search_nodes

`query`, `category`

auto_fix_workflow

`workflow_json`, `workflow_name`

get_workflow_guide

`use_case`

T-04

Auth — Wrong Token

T-11

Auto-Fix

T-18

Rate Limiting

auto_stop_machines

`false`

Task

Frequency

M-01

Core MCP server with FastAPI + FastMCP