This automation enables testing and improving AI agent logic for handling user queries via chat or structured datasets. The system uses multiple agents and tools (knowledge base search, web search, calculator), evaluates action accuracy, and logs metrics to Google Sheets. Ideal for teams building multi-step LangChain systems within n8n.
## Who it´s for
- AI agent developers testing tool call correctness
- MLOps engineers evaluating agent performance
- Teams building multi-step LangChain systems in n8n
- Analysts comparing model outputs using structured datasets
## What the automation does
- Processes input from chat messages or Google Sheets rows
- Triggers a multi-agent system using OpenRouter and LangChain
- Performs knowledge base search (Qdrant), web search (Firecrawl), math calculations, and summarization
- Evaluates whether correct tools were selected and used, based on expected values from Sheets
- Logs execution results, metrics, and evaluations back to Google Sheets
## What´s included
- Ready-to-use n8n workflow with multi-agent orchestration
- Trigger logic for chat messages or new Google Sheets entries
- Integrations with Qdrant, Firecrawl, Google Sheets API, OpenRouter, and OpenAI
- Basic setup and adaptation guide
## Requirements for setup
- n8n instance with HTTP API access and workflow management
- Google Sheets account with read/write permissions
- Qdrant account for vector storage
- OpenRouter API key (for LLM inference)
- OpenAI API key (for embeddings)
- Firecrawl API key (for web search)
## Benefits and outcomes
- Automated agent logic testing without manual oversight
- Ability to compare different models’ effectiveness via unified dataset
- Transparent evaluation of tool usage accuracy
- Centralized logging of results and metrics in Google Sheets
- Faster iteration and improvement of agent behavior
- Support for complex workflows: hybrid search, calculations, analysis
## Important: template only
Important: you are purchasing a ready-made automation workflow template only. Rollout into your infrastructure, connecting specific accounts and services, 1:1 setup help, custom adjustments for non-standard stacks and any consulting support are provided as a separate paid service at an individual rate. To discuss custom work or 1:1 help, contact via Telegram: @gleb923.
multi-agent system
chat query processing
tool usage evaluation
AI agent testing
n8n workflow
LangChain agents
Google Sheets evaluation
knowledge base search
web search Firecrawl
semantic search Qdrant
OpenRouter LLM
OpenAI embeddings
automated query processing
tool calling logic
agent performance analysis
multi-step workflow
No feedback yet