Build AI workflows that connect to the tools your team already uses. Configure integrations visually, compose multi-step pipelines, and monitor every execution.
From configuration to monitoring — see how a real workflow connects AI models, databases, and communication tools.
1
Choose Your Integrations
Browse the catalog and connect the tools you need. Authenticate with OAuth, paste an API key, or point to an MCP server.
2
Compose Your Workflow
Drag integrations onto the visual canvas. Wire triggers, decisions, and actions into an execution graph.
3
Configure Logic
Add routing rules, payload transforms, and error handling. Every path is explicit and inspectable.
4
Monitor & Iterate
See every run in real time. Trace failures to the exact step, retry with one click, and set up alerts.
Search integrations...⌘K
SlackOAuth
OpenAIAPI Key
PostgreSQLMCP
GitHubOAuth
TwilioAPI Key
Drag to canvas to add · 5 shown
1Choose Your Integrations
Browse the catalog and connect the tools you need. Authenticate with OAuth, paste an API key, or point to an MCP server.
Search integrations...⌘K
SlackOAuth
OpenAIAPI Key
PostgreSQLMCP
GitHubOAuth
TwilioAPI Key
Drag to canvas to add · 5 shown
Three Ways to Connect
Every integration uses one of three authentication methods. Choose the right one for your use case.
MCP Servers
Connect through Model Context Protocol servers for structured tool access. MCP gives your workflows direct, typed access to databases, file systems, and custom tools with explicit permissions.
Typed tool interfaces
Scoped permissions
Local or remote servers
Used by:
OAuth
Authenticate with user-scoped tokens from SaaS platforms. OAuth connections handle token refresh, scope management, and revocation automatically.
Automatic token refresh
Granular scope control
One-click revocation
Used by:
API Keys
Bring your own credentials for any API. Store keys securely in your workspace vault and reference them in workflows without exposing secrets.
Encrypted vault storage
Per-workflow key rotation
Usage tracking
Used by:
Get Connected in Minutes
Three steps from zero to a production-ready integration workflow.
1
Authenticate
Connect with OAuth, paste an API key, or point to an MCP server. Credentials are encrypted and scoped to your workspace.
2
Build Your Workflow
Drag integrations onto the visual canvas. Wire triggers, decisions, and actions into a deterministic execution graph.
3
Deploy & Monitor
Push to production with one click. Every run logs inputs, outputs, latency, and cost across every integration call.
Integration FAQs
Common questions about connecting tools and services to NodeFox.
What is MCP and how does it differ from a REST API integration?
MCP (Model Context Protocol) provides a structured, typed interface for tools. Unlike REST APIs where you configure HTTP endpoints manually, MCP servers expose discoverable tools with defined input/output schemas. This means your workflows get auto-complete, validation, and permission scoping out of the box.
Can I connect a tool that isn't in the pre-built catalog?
Yes. The REST API and GraphQL integrations let you connect to any HTTP endpoint. For more structured access, you can run your own MCP server and point NodeFox at it. Custom integrations get the same monitoring, retry, and governance features as pre-built ones.
How are API keys and OAuth tokens stored?
All credentials are encrypted at rest in your workspace vault. OAuth tokens are automatically refreshed before expiry. API keys are never exposed in workflow logs or execution traces -- they are injected at runtime and masked in all UI surfaces.
Can one workflow use multiple authentication methods?
Absolutely. A single workflow can mix MCP tools, OAuth-connected SaaS apps, and API-key services. Each integration node in the graph manages its own authentication independently.
What happens when an integration fails mid-workflow?
NodeFox provides configurable retry policies, exponential backoff, and fallback branches. Every failure is logged with the full request/response context so you can diagnose issues without guessing. You can set up automatic escalation to Slack or email on repeated failures.
Is there a limit on the number of integrations per workflow?
No hard limit. Workflows can connect as many integrations as your use case requires. Each integration call is independently monitored for latency and cost, so you maintain full observability as workflows grow in complexity.
Start Building Integration Workflows
Connect your first integration in under five minutes. No credit card required.