Testing Overview
- Time spent: ~2 hours
- Workflow built: Webhook → AI Agent → Conditional Branching → HTTP Actions
- Focus: General usability, AI Agent functionality, execution debugging
What Worked Well
Drag-and-drop interface - intuitive node placement and connection
AI Agent node - successfully classified messages without requiring API keys (impressive!)
Data picker - easy to reference previous step outputs
Workflow activation via “Live” toggle works once discovered
Issues Found
1. AI Agent Executions Tab Shows “No Recent Data”
Description: AI Agent processes successfully (green checkmark), but Executions tab displays “No Recent Data”
Expected: Should show execution history with inputs/outputs for debugging
Steps to reproduce:
- Create workflow: Webhook → AI Agent → Condition
- Send POST request to webhook
- AI Agent executes (green checkmark appears)
- Click AI Agent → Executions tab
- Result: “No Recent Data”
Impact: Cannot debug AI Agent behavior or verify classification output
2. Unclear Workflow Activation Process
Description: Workflow shows “Triggers inactive. Manual runs only.” but no obvious way to activate
What I found: “Live” toggle exists next to workflow name (top left) but is easy to miss
Suggestion: Add more prominent “Activate Workflow” button or banner when triggers are inactive
3. Auto-Created HTTP Nodes Cannot Be Configured
Description: When adding action to Condition’s “Then go to” branch, platform auto-creates “Send POST Request” node but configuration panel doesn’t load or is inaccessible
Workaround: Delete auto-created node and manually add HTTP node via canvas
Impact: UX friction when building conditional workflows
4. Condition Configuration Resets When Adding Actions
Description: After configuring IF condition, clicking “Select step” to add action clears the IF configuration
Impact: Must re-configure condition after adding actions
5. Limited Execution Trace Visibility
Description: Unable to see full workflow execution trace across all nodes in single view
Expected: Unified execution view showing data flow through all steps (like n8n’s execution view)
Current: Each node’s Executions tab shows isolated data
Impact: Difficult to debug multi-step workflows or verify branching logic
Questions
- Is there a way to view full execution traces across all nodes?
- Is the AI Agent Executions tab issue a known bug?
- Are there plans to improve trigger activation UX?
- What’s the recommended way to debug conditional branching when execution data isn’t visible?
Overall Impression
GetGlow has strong potential, especially the AI Agent with built-in MCP tools. Main pain points are around execution visibility and debugging multi-step flows. With better execution tracing and clearer activation controls, this could be very competitive.
Happy to provide more details or test specific scenarios.