Workflow automation with workflow memory
Automate once, remember every run
The Problem
n8n makes workflow automation visual and powerful. Drag, drop, connect, automate. HTTP requests, database queries, API integrations, webhooks—all connected in a visual canvas. Your workflows run on schedule, handle events, process data. The open-source alternative to Zapier that developers actually love.
But each run starts from scratch.
Here's the limitation: Your n8n workflow processes customer support tickets. Today it categorizes, routes, and responds to 500 tickets. Works perfectly. It encounters a weird edge case—a ticket with unusual formatting that needs special handling. You add logic to handle it. Tomorrow, 500 more tickets. But your workflow doesn't remember yesterday. That edge case handling? Applied to yesterday's data, but the insight isn't available for today's analysis.
The problem compounds over time. Your workflow has processed millions of records. It's seen every pattern, handled every edge case, recovered from every error—and learned nothing. Each execution starts fresh, with no awareness of what came before. The workflow itself improves (you update it), but the execution context doesn't persist.
For AI-powered n8n workflows, this is especially limiting. Your AI nodes make decisions based on patterns, but those patterns are discovered and forgotten in a single run. The next run makes the same discoveries. Expensive pattern recognition repeated infinitely.
Automation without memory is just scheduled amnesia.
How Stompy Helps
Stompy gives your n8n workflows persistent memory that compounds across executions.
Your automated workflows gain true learning: - **Execution memory**: Each run knows what happened in previous runs. Edge cases handled yesterday inform handling today. - **Pattern accumulation**: Patterns discovered during processing are saved and available to future executions. Your AI nodes get smarter over time. - **Error wisdom**: When a workflow recovers from an error, the recovery strategy is remembered. Same error next time? Already solved. - **Cross-run context**: Data transformations, categorization decisions, routing patterns—all persist and inform future runs.
Your visual workflow design stays in n8n. Your execution intelligence stays in Stompy. Together: workflows that genuinely learn from every run.
Workflows that learn from every execution.
Integration Walkthrough
Create Stompy memory nodes for your workflow
Use n8n's HTTP Request node to connect to Stompy for context recall and storage.
{"nodes": [{"name": "Recall Past Patterns","type": "n8n-nodes-base.httpRequest","parameters": {"url": "https://mcp.stompy.ai/sse","method": "POST","headers": {"Authorization": "Bearer {{ $env.STOMPY_TOKEN }}","Content-Type": "application/json"},"body": {"tool": "context_search","query": "{{ $json.category }} patterns","limit": 5}}},{"name": "Save New Insight","type": "n8n-nodes-base.httpRequest","parameters": {"url": "https://mcp.stompy.ai/sse","method": "POST","headers": {"Authorization": "Bearer {{ $env.STOMPY_TOKEN }}","Content-Type": "application/json"},"body": {"tool": "lock_context","topic": "workflow_{{ $workflow.name }}_insight_{{ $now.format('YYYYMMDD') }}","content": "{{ $json.insight }}","priority": "reference"}}}]}
Build a learning data processing workflow
Create a workflow that recalls past patterns before processing and saves new insights after.
{"name": "Customer Support Classifier with Memory","nodes": [{"name": "Webhook Trigger","type": "n8n-nodes-base.webhook","position": [250, 300]},{"name": "Recall Category Patterns","type": "n8n-nodes-base.httpRequest","position": [450, 300],"parameters": {"url": "https://mcp.stompy.ai/sse","method": "POST","body": {"tool": "recall_context","topic": "support_categorization_patterns"}}},{"name": "AI Classifier","type": "@n8n/n8n-nodes-langchain.openAi","position": [650, 300],"parameters": {"prompt": "Categorize this ticket using these known patterns:\n{{ $node['Recall Category Patterns'].json.content }}\n\nTicket: {{ $node['Webhook Trigger'].json.text }}"}},{"name": "Save If New Pattern","type": "n8n-nodes-base.if","position": [850, 300],"parameters": {"conditions": {"string": [{"value1": "={{ $json.is_new_pattern }}", "value2": "true"}]}}}],"connections": {"Webhook Trigger": {"main": [[{"node": "Recall Category Patterns"}]]},"Recall Category Patterns": {"main": [[{"node": "AI Classifier"}]]},"AI Classifier": {"main": [[{"node": "Save If New Pattern"}]]}}}
Track execution metrics for workflow optimization
Log execution patterns to identify optimization opportunities across runs.
{"name": "Execution Logger","type": "n8n-nodes-base.httpRequest","position": [1050, 300],"parameters": {"url": "https://mcp.stompy.ai/sse","method": "POST","body": {"tool": "lock_context","topic": "workflow_{{ $workflow.name }}_metrics","content": "Execution: {{ $now }}\nRecords processed: {{ $items.length }}\nCategories: {{ $json.categories.join(', ') }}\nErrors: {{ $json.error_count }}\nDuration: {{ $execution.duration }}ms","tags": "metrics,workflow","priority": "reference"}}}// Query past metrics for optimization:// Use context_search with "workflow metrics errors" to find problem patterns// Use context_search with "workflow duration slow" to find bottlenecks
What You Get
- Cross-run learning: Patterns discovered in execution #1 inform execution #1000, creating genuinely improving workflows
- Error recovery memory: When your workflow solves a problem, the solution is available next time without re-debugging
- Visual design + persistent intelligence: Keep n8n's beautiful visual workflow builder while adding execution memory
- AI node enhancement: Your AI-powered nodes get smarter over time as patterns and decisions accumulate
- Execution analytics: Search past runs semantically to identify optimization opportunities and recurring issues
Ready to give n8n a memory?
Join the waitlist and be the first to know when Stompy is ready. Your n8n projects will never forget again.