I am pleased to annouce my little and successful work with my Rayban Meta + Tasker to convert your glasses into a ChatGPT vision with your own language. (my hometown is Hong Kong, ChatGPT & Meta AI has blocked me)
materials:
- ChatGPT API
- Tasker with ChatGPT in Taskernet + Autonotification + Autovoice
- a bit modification for ChatGPT plugin
- Whatsapp & Whatsapp Business A/C on same phone
- VPN account
sorry i am not familiar on how to upload my tasks here, i try to upload my Task screenshot to you first and if anyone interested in it just inbox me / expertees to share for me
i. Use Rayban Meta to take photos, send Whatsapp to your own W4B account
ii. Autonotification intercept message from Whatsapp Business
iii. Trigger tasker to run task, connect VPN (disconnet afterwards to save battery)
iv. run shell to locate the latest W4B image and copy to new path
v Varaibleset to perform task from ChatGPT plugin
vi Pick the saved photos
vii and now you can use ChatGPT to ask question in your languages if you have set to your own
viii response the answer with Text to speech and listen from your Rayban Meta
I really love this hardware and woule see if this community can strenthen the fewatures of this Glasses, let’s chat and optimize together, cheer!
Alan is genius. This is exactly what I need to work around whenever Meta tells me it can’t read something. I am legally blind and depend on my glasses. Unfortunately, I am no longer proficient with some of the apps needed.
Nice, do you have a tutorial? Would love to test it.
Very nice Alan. To share the tasker apps you to long hold the profile then share. There will be an option to share/upload to taker for a link to share.
Once you have done that. You can post the link here for others to get.
just made a very simply demo video in Cantonese version, i believe not much people understand
however just want to present this Rayban Meta thru ChatGPT can now use its camera to make things happened
Profile: Take photos from Glasses & send to Whatsapp Business
Event: AutoNotification Intercept [ Configuration:Event Behaviour: true
Notification Type: Only Created Notifications
Has Reply Action: true
Notification Apps: WhatsApp Business
Notification Title: Your contact name
Notification Text: Image ]
Enter Task: Enable ChatGPT action
A1: Perform Task [
Name: OpenVPN connect
Priority: %priority
Structure Output (JSON, etc): On ]
A2: Variable Set [
Name: %path
To: Android/media/com.whatsapp.w4b/WhatsApp Business/Media/Whatsapp Business Images/Private/
Structure Output (JSON, etc): On ]
A3: Run Shell [
Command: ls "/sdcard/%path" -at | head -n1
Timeout (Seconds): 0
Store Output In: %File_name
Store Errors In: %File_error
Use Global Namespace: On ]
A4: Copy File [
From: %path%File_name
To: Tasker/meta.jpg
Use Global Namespace: On ]
A5: Perform Task [
Name: Meta Question image
Priority: %priority
Structure Output (JSON, etc): On ]
A1 is the VPN connect task if necessary
A5 is the modified task that use Tasker ChatGPT plugin (Pick and question image) that part
2nd day
i made some tricks to modify the Tasker step
instead of running shell to load the last received photo, i change the location to last Sent photo from Whatsapp folder
and turning off Whatsapp business auto media download function
in order to save one time of the download data and with a bit faster loading speed
Meta AI is no longer available in hk,…
Hey , can und make a video tutorial about this? Step by Step 
thank u!!!
How do you put this into the glasses exactly?
i am not programmer too and just enthusiastic, i do wish people around here could be inspired by my solutions and build a better ways because of my poor programming skills and english…
I also want to share to all, however don’t know how to…
just knew how to do it, thanks a lot
VPN is only if API access is geoblocked, right?
A5 part
Task: Meta Question image
A1: Say [
Text: 想問咩？
Engine:Voice: com.google.android.tts:yue-hkg
Stream: 3
Pitch: 5
Speed: 5
Respect Audio Focus: On ]
A2: AutoVoice Recognize [
Configuration: Hide Dialog: true
Voice command with headset
Command: "all"
Timeout (Seconds): 120
Structure Output (JSON, etc): On ]
A3: Progress Dialog [
Action: Show/Update
Title: image Recognition
Text: Uploading image...
Type: Animation
Frame Duration: 66
Progress: 0
Max: 100 ]
A4: Perform Task [
Name: System >> Upload And Share Image
Priority: %priority
Parameter 1 (%par1): Tasker/meta.jpg
Return Value Variable: %url ]
<Sometimes the image upload service will serve an old version of the image, do wait a second to give it time to update.>
A5: Wait [
MS: 0
Seconds: 1
Minutes: 0
Hours: 0
Days: 0 ]
A6: Progress Dialog [
Action: Show/Update
Title: image Recognition
Text: Analyzing image...
Type: Animation
Frame Duration: 66
Progress: 0
Max: 100 ]
A7: Perform Task [
Name: System >> Clear Chat
Priority: %priority
Structure Output (JSON, etc): On ]
A8: Perform Task [
Name: System >> Send Chat
Priority: %priority
Parameter 1 (%par1): %avcomm
Parameter 2 (%par2): %url
Return Value Variable: %response
Structure Output (JSON, etc): On ]
A9: Progress Dialog [
Action: Dismiss
Type: Animation
Frame Duration: 66
Progress: 0
Max: 100 ]
A10: Open File [
File: Tasker/meta.jpg
Mime Type: image/* ]
A11: Perform Task [
Name: System >> Say Response
Priority: %priority
Parameter 1 (%par1): %response
Structure Output (JSON, etc): On ]
I am sorry that I couldn’t
Because I don’t have enough time to making video
And I am still rebuilding it
I only get a download up to step a5, is there more after?
in taskernet, search chatgpt, download and modify it
thank u!!!