Error workflow: AI powered (GPT 4.1): universal
Needs changesNeeds changes. This template uses 1 thing(s) that n8n 3.0 removes: AI Agent v1 (old agent modes). 1 breaking item(s) found. Static check, last run 2026-10-19.
What this workflow does
This AI workflow has 6 nodes and runs when another workflow fails (error workflow). Along the way it connects OpenAI Chat Model, Gmail and Telegram.
How it runs
Trigger: When another workflow fails (error workflow).
- AI Agent: runs an AI agent, with model: OpenAI Chat Model
- Format Data (Edit Fields (Set)): sets 8 fields: errorMessage, nodeName, nodeType…
- Send Gmail Notification (Gmail): uses Gmail
- Text Notification (Telegram): uses Telegram
Services it uses
Before you upgrade to n8n 3.0
1 thing in this workflow is removed in n8n 3.0. Fix it before you upgrade:
- AI Agent v1 (old agent modes) in AI Agent (v1.7): replace with Latest AI Agent node (Tools Agent). For SQL Agent, use a Postgres or MySQL tool with a recent AI Agent.
Rather not rewrite it? Many of these flows can be rebuilt in Make (free plan).
| Removed in 3.0 | Replace with |
|---|---|
| AI Agent v1 (old agent modes) AI Agent (v1.7) | Latest AI Agent node (Tools Agent). For SQL Agent, use a Postgres or MySQL tool with a recent AI Agent. |
Nodes used (6)
- Error Trigger
n8n-nodes-base.errorTrigger - AI Agent
@n8n/n8n-nodes-langchain.agent - OpenAI Chat Model
@n8n/n8n-nodes-langchain.lmChatOpenAi - Edit Fields (Set)
n8n-nodes-base.set - Gmail
n8n-nodes-base.gmail - Telegram
n8n-nodes-base.telegram
From the author
AI-Powered n8n Error Debugger & Notifier Automatically analyze any workflow failure with AI, get actionable solutions, and receive a detailed report directly in your inbox. Stop wasting time deciphering cryptic error messages and stack traces. This template…