Query OpenAI Large Language Models via Email

Joke or not, this remains the most reliable way to authenticate the sender of an email.

Exactly the most reliable.

In the same way as when you jump into a lake pressing a hand on your eyes is the most reliable way to keep your body dry.

I mean there must be a reason why worldcoin scans your eyes…(other than making it possible to send migrants back to where they came from, when they “lost” their passport after illegal immigration)

Well Digital Ocean seems to have locked down any ports other than 22, 443, 80

Can’t even get there from sendgrid anymore

Support is stalling

Seems like spam has destroyed inbound mail for the outdoor dogs

What is your program logic?

This is what is working for me:

user sends email —> email is forwarded to Zapier —> Zapier Email Parser —> Zapier Webhook —> My AI Webhook —> Question submitted to LLM (API) —> LLM returns response —> My AI Webhook sends response back —> sender email

Works perfectly.

Or ask chatgpt to write you a short php script and run it with a cron job. Low code is dead haha…

https://chat.openai.com/share/fd5e3530-66da-492d-ae09-1f2f46956303

I mean… 10 minutes more should be invested…

1 Like

I have my :fox_face: reading all my mail now i forwarded to root very nice

Now to reply

So it is possible now to CC: GPT and get a email reply

:confetti_ball:

Sendgrid has a webhook call I listen for and get all the mail for talkeetna.info in one file I decode and feed it into the :fox_face: and ask him for a summary and a email

2 Likes

1 Like

The SMS responder was relatively easy to set up, and working fairly well.



Using only Twilio SMS API and Drupal Webbook on my site.

1 Like

Hi, do you have a setup for this? I am trying to integrate via Replit for Now. The libraries are causing so many errors in the Configuration Line and the Create lines. All the docs were confusing and even tried to combine to working replits.

If you mean the SMS responder, yes, it’s working. I communicate directly between the APIs.

User SMS Phone → AI SMS Phone → Webhook → AI ChatCompletion → Webhook → AI SMS Phone → User SMS Phone

So, I have an AI SMS Phone set up to receive/ send texts. It receives text from end-user, sends it to my application Webhook which sends it as a chat completion query in my application. The application responds to the Webhook, which sends the response back to the AI SMS Phone which texts the response back to the end-user.

That’s pretty much the process. I don’t know what Replit is, and I don’t use libraries, but I’m happy to answer any questions you might have about how I know to do it.