@Josh Wardini yes it’s not an easy template to use. Lots of setup.

I have an n8n up and running on Railway and it seems to be functioning properly for 6+ months with active workflows. Can you provide a few examples of your workflows that are Active and what the triggers look like and what the settings are. Its hard to give you any practical advice without more information.

looks good. all I can think of is that your gmail credentials are not setup correctly or your poll time is more than Every minute. have you also checked your Executions to see if there is anything else causing the workflow to potentially fail. Sometimes you just need to create new credentials and test them by sending an email from n8n to ensure they work before using on the Trigger node.

Another approach to using Telegram with n8n is using webhooks instead of Telegram triggers. This workflow shows that approach with the Telegram Webhook tools to adjust from testing to production. Maybe this approach will work for you too: https://n8n.io/workflows/2751-telegram-messaging-agent-for-textaudioimages/

@Stjepan Vidovic Great you found a solution. I originally tried Render and quickly switched to Railway. Railway has been great and very reliable <$10 per month.

Hi Nathan. I don't have any videos but I do have a few workflows that use those nodes a lot. Happy to share those links to workflows if that would help you to see those nodes in the context of a larger workflow.

@Nathan Weilbacher You can browse my community n8n workflows here and there are many examples of those node in use. https://n8n.io/creators/joe/. This workflow uses merge, split and the code nodes which are really handy. https://n8n.io/workflows/3003-monitor-favorite-youtube-channels-through-rss-feeds-and-receive-notifications/ This workflow uses a loop and splits a list: https://n8n.io/workflows/3185-compare-local-ollama-vision-models-for-image-analysis-using-google-docs/

@Imad L I built a voice agent with Python and n8n that runs on my local server and dev machine. I interact with it like a Siri using microphone. It also has access to my computer via Python and my other n8n workflows for Gmail, Google docs, etc. via webhooks. My solution was built for myself as a proof of concept and not a pubic facing app.

@Imad L It started here with Sam Witteveen YouTube about Kokoro Loal TTS https://www.youtube.com/watch?v=tl1wvZXlj0I and from there I used Python to build a FastAPI using Claude 3.5 and integrated with n8n web hooks and HTTP calls back the to FastAPI endpoints. Unfortunately I do not have any videos or docs for that part of the app. Sam does a very good job of explaining Kokoro and shared a Google Colab to get you started. https://colab.research.google.com/drive/14qaWVuAO37lCLbqhHAMApC6825wvZeJ7?usp=sharing