Buildship Meditation GPT
Hello. My name is Jason. I am trying to learn and use Buildship to Flutterflow to build a voice-first/chat assistant that combines various somatic wellness techniques to create routines to help be enhance their wellness. Coming from a background in web design and digital marketing, as well as having a certification in breathwork, I wanted to come up with something that would combine my many skills and interests. I have for the most part been enthusiastic about learning new skills and tools.
I got as far as planning and developing the concept and UI in Flutterflow. I'm finally starting to get my head around how to build the front-end in Flutterflow and connecting it all together. Where I am struggling most is in Buildship.
It took me a month, but I finally, after much persistence, figured out how to get how to get streaming chat to work. Now I am trying to tackle the next level in my Buildship learning... the Meditation GPT template.
Initially, started to set it up with an ElevenLabs voice I found, and liked very much, but as I ran into a snag in testing, I watched as my character count (each test) dwindled down by 1000. I saw in the logs that the text-speech was working, but where it's not is in the "Combine Audio" node, I am consistently getting a "FFmpeg exited with code 1" error and I have no idea why. I copied the template from the Template library, and followed the video on YouTube, step-by-step.
Though, I did change the text-to-speech from ElevenLabs to Open AI's tts-1-hd...
One thing I noticed is that in the video it shows a JSON Parse node, but that node isn't in the template. The video and the template do not seem to be one-to-one...
I've been slow to reach out and ask anyone for help, trying to read the documentation, watch the videos, and learn on my own. I used to think I was a pretty smart guy and could figure these things out. Now I'm not so sure anymore. lol
If anyone could offer some support, I would be most grateful. Thank you, Jason
Solution:Jump to solution
Hi @Jason Samadhi, Sent you an email with cause of the issue and solution, you can reply all to the email for any further questions.
Also, sharing the reply here - Regarding the error "FFmpeg exited with code 1" on the combine audio node, on taking a look at your modified workflow (iffvISDRGkXPfXo56kyk), you have switched the playHT node with OpenAI Text-to-Speech node, which is causing the issue here. Elaborating the issue and the solution below:
Issue - The OpenAI Text to Speech node returns a base64 instead of URL whereas the combine audio node only accepts an array of URL as input in the audio file path....
BuildShip - Visual backend workflow builder
Visually build workflows and powerful backend logic for your apps. Powered by AI, connect to anything with or without code.
6 Replies
Hi Jason,
It sounds like you're making great progress with your project! Let's address the issues you're encountering with the Meditation GPT template.
FFmpeg Error
The "FFmpeg exited with code 1" error typically indicates that there is an issue with the command or the input files used in the FFmpeg process. Here are a few steps you can take to troubleshoot this:
1. Check Input Files: Ensure that the audio files being combined are in the correct format and are accessible.
2. Command Syntax: Verify that the FFmpeg command in the "Combine Audio" node is correct and properly formatted.
3. Logs: Review the detailed logs for any additional error messages that might give more insight into the issue.
JSON Parse Node
It's possible that the template or the video may have been updated or changed. If the JSON Parse node is missing from the template, you might need to manually add it to ensure that the workflow functions correctly.
Text-to-Speech Service
Switching from ElevenLabs to Open AI's tts-1-hd might also require some adjustments. Make sure the output format from the text-to-speech service is compatible with the rest of the workflow.
Documentation and Support
For more detailed guidance, you can refer to the official BuildShip documentation. Here are some relevant links:
- WhatsApp Bot Documentation
- Generate EPUB Node Documentation
If you continue to face issues, I recommend submitting a support request via the in-app support button. The BuildShip team can provide more personalized assistance to resolve your specific problems.
Keep up the great work, and don't hesitate to reach out if you need further assistance!
Best regards,
[Your Name]
WhatsApp Bot – BuildShip
The WhatsApp Bot is a powerful trigger that allows you to connect and create a WhatsApp Bot using using WhatsApp Business Account by Meta.
Generate EPUB Node 📖 – BuildShip
A unified resource to start building your backend with low-code. Dive into triggers, nodes, and step-by-step guidance to jumpstart your workflow creation.
Hey @Jason Samadhi, love your idea. This is true that template has been updated a bit since the video was released. Can you submit an in-app support request so I can further assist you with your workflow?
Sure! If you have any questions or need further assistance with your workflow, please submit an in-app support request via the BuildShip app. This will allow our team to help you more effectively.
I have. As well as several comments regarding this issue plus some issues I am having with your new Streaming API Response template for Flutterflow. I followed the instructions, again step-by-step, and I couldn't get it to work either. In addition, it doesn't make sense to me. Why would you save the chatHistory in the page state? Wouldn't you want the chatHistory saved in the app state, and for that matter in the database? It feels like there are steps in this video missing, just as there are inconsistencies in the Meditation GPT app.
For your BuildShip question and the issues you're experiencing with the Streaming API Response template for Flutterflow, I recommend submitting your detailed queries via the in-app support button. This will allow the support team to review your specific use case and provide tailored assistance.
Regarding saving the chat history, it indeed makes sense to save it in a persistent storage like a database for long-term access and consistency across sessions. You can refer to the following documentation for more information on how to save chat history to a database:
- Saving Chat History Externally in a Database
For more detailed guidance and troubleshooting, please submit your issues via the in-app support button.
AI Assistant Builder Guide – BuildShip
With BuildShip's AI Assistant Builder Integration, you can now leverage the powerful capabilities of AI Assistants without writing any code visually.
Solution
Hi @Jason Samadhi, Sent you an email with cause of the issue and solution, you can reply all to the email for any further questions.
Also, sharing the reply here - Regarding the error "FFmpeg exited with code 1" on the combine audio node, on taking a look at your modified workflow (iffvISDRGkXPfXo56kyk), you have switched the playHT node with OpenAI Text-to-Speech node, which is causing the issue here. Elaborating the issue and the solution below:
Issue - The OpenAI Text to Speech node returns a base64 instead of URL whereas the combine audio node only accepts an array of URL as input in the audio file path.
Solution - Using any other text-to-speech node that return a URL should fix the issue. For example, I used Eleven Labs text-to-speech node and the new combine FFmpeg node from the FFmpeg integration, which worked perfectly fine.
For your reference and ease, I've created a modified working and tested workflow for Meditate GPT which you can clone and use by adding your api-keys/secrets. Also attaching a test screenshot and a remix link of the workflow.
Remix Link - https://buildship.app/remix/1166d4de-cc43-494b-84f4-fdafd6f031d4.
Hope this helps and keeps you going with BuildShip. 🙂
BuildShip - Visual backend workflow builder
Visually build workflows and powerful backend logic for your apps. Powered by AI, connect to anything with or without code.