Forum n8n Community

I had a similar issue with a cloud service before, and in my case the account payment was fine but the workspace itself was having problems. Since you were able to restart it through the magic link, I would check the workspace status and billing area as well, just to make sure there isn’t a separate account or subscription issue. Hopefully the restart gets your workflows running again.

Hi @vfavrat , your webm node also has On Error set to Continue. An SMTP failure can then appear as output data instead of stopping the workflow. Test a copy in a separate workflow: Manual Trigger → Send and Wait, with the same SMTP credentials, your own recipient address and a short fixed subject/message. Leave Grist disconnected. In that copy, set Settings → On Error → Stop Workflow, unpin the node and move the wait deadline into the future. A failed send should now show the actual error. If that isolated node reaches Waiting but nothing arrives, check the SMTP server’s delivery logs and quarantine. Waiting means the SMTP send completed, not that the email reached the inbox.

Hi, I’m interested in the B2B lead generation POC. My strongest hands-on areas are n8n, APIs/webhooks, AI/LLM workflows, data processing, JavaScript and business automation. I’m particularly interested in the lead discovery → enrichment → qualification → personalized outreach pipeline you described. I want to be transparent that I would not claim deep production LangGraph experience. My stronger side is the n8n/API/LLM orchestration layer. For a first milestone, I’d suggest building a clearly testable slice around official API lead discovery, normalization/enrichment, qualification, CRM handoff and human approval before any outreach. I can deliver the workflow as exportable JSON with documentation, validation and failure handling, and I’m comfortable working within a fixed-price POC. If you’re still reviewing candidates, I’d be happy to discuss the first milestone and show my existing n8n/AI automation work. Best, Vahid

Many thanks for making learning resources easily available to beginners like myself. The concepts are explained clearly!

Hi Ahmed, I’m interested in the long-term technical automation role. My strongest hands-on areas are n8n, REST APIs, webhooks, LLM integrations, structured AI workflows, data processing, JavaScript and workflow reliability. I’m particularly interested in the n8n/API/CRM/lead qualification side of your systems. I haven’t shipped a production Retell/Vapi/Twilio voice system yet, so I don’t want to overstate that experience, but I’m comfortable with API-based integrations and can learn the platform-specific layer quickly. I’d suggest starting with a small paid n8n/API/CRM workflow or repair task. That gives you a direct way to evaluate how I build, test, document and handle failure cases before moving into larger client systems. I’m available for ongoing project-based work and can start immediately. Best, Vahid

Hi, I’m interested in your Tally → Airtable → n8n automation project. The workflow you described is a good fit for the type of automation I build: structured lead capture, approval checkpoints, state-based logic and automated workflow control. I would structure it around clear states so that: Tally submission → Airtable record → Approval 1 → Approval 2 → downstream action and separately use the payment status as a hard guard so overdue clients cannot continue triggering the workflow. I’d keep the workflow modular and documented so you can understand and manage it yourself after handover. I’m comfortable working on a fixed-price basis. If you send me the full brief, I can review the complete scope and give you a clear fixed quote and delivery time. I can start immediately. Best, Vahid

Hi Colin, I’m interested in the Business Performance Analyzer project. I want to be transparent about my current experience: I have strong hands-on experience with n8n workflow design, APIs/webhooks, structured data processing, AI/LLM workflows, validation, error handling and automation architecture, but I have not yet shipped this exact Amazon Ads + Keepa + ClickUp system in production. Rather than pretending otherwise, I’d suggest starting with a tightly scoped Phase 1. I would focus first on the Ad Waste Detector: Amazon Ads data → normalized metrics → deterministic ACoS/spend rules → Supabase snapshot → traceable Slack alerts, with duplicate protection, retries and visible failure handling. Claude can then explain and rank the results, while the actual financial calculations remain deterministic and traceable. I’m comfortable working from your existing architecture, using the provided API access, and delivering the workflow with documentation and testing. I’d be happy to discuss a clearly defined first milestone and prove the fit before moving into the larger build. Best, Vahid

Hi, I’m interested in the first phase of your household CTO project. My strongest hands-on area is n8n automation, API/webhook integrations, AI/LLM workflows, Google Workspace, structured data handling, workflow logic, testing, error handling and documentation. I have not deployed this exact Mac mini + Ollama/LM Studio household setup in production before, so I don’t want to overstate that experience. What I can confidently handle is the n8n orchestration and the integrations around it. For the first $1,000–$2,000 phase, I’d keep the scope focused: define the local-vs-cloud data flow, build one useful end-to-end workflow with human approval where appropriate, test failure cases, and provide a simple runbook so the system is maintainable. I’m comfortable working remotely and asynchronously, and I’m happy to start with a clearly defined first milestone. Best, Vahid

This should help Docker Ollama not being able to call tools Questions Hi @sneakylenny Welcome! Docker Model Runner’s Ollama-compatible endpoints are a thin shim and tool calls don’t survive it. Your capture shows exactly that: finish_reason comes back as tool_calls but the message carries no tool_calls array, so the agent gets an empty assistant turn and returns { "output": "" }. Docker documents function calling only on its OpenAI-compatible surface, which is why all five models behaved identically. Swap the Ollama Chat Model node for the OpenAI Chat Model node… You need to create an OpenAI credential using Ollama base url

i tried but i can not put my ollama inside the chat gtp credential type, maybe i am missing something but the wierd thing is that in your image, you have ollama with your credential i think but at least the provieder and i am not, there is something that i am missing like a configuraction or something in the new agents (previed) section?

Docs PR is up: docs: Document config-as-code provisioning for external secrets by kourosh-forti-hands · Pull Request #5486 · n8n-io/n8n-docs · GitHub It’s in draft until the code PR ( #39655 ) merges and I know the real version number to put in the “available from n8n X.Y.Z” note.

I think you can do it this way