New project: clank.email

Here's a new thing from me: clank.email.

This lets you send an email to an address (yo@clank.email) and get an LLM response back.

There have been a number of times when I received an email that I wanted to process with an LLM for various reasons. For example, I once sent myself a preview of a newsletter draft and wanted an LLM to check it for mistakes. Another time, I opened a very long thread I hadn't been keeping up with and needed a quick summary. It's somewhat cumbersome to expand the entire thread, copy the contents, paste them into ChatGPT, get a response, and then bring it back to the email. Clank is designed to make LLMs accessible directly in-thread.

I imagine there are many other use cases for bringing LLM functionality into email threads, which is why I built this with a reasonably generous free tier (the first 5 emails are free). Based on early usage, forwarding emails seems to be the most common workflow.

I set up basic paid plans: $6/month for up to 200 responses, and $30/month for effectively unlimited responses. I plan to expand the unlimited tier to allow custom instructions, multiple originating email addresses, and access to better models. For now, these plans exist (1) to test demand and (2) to impose usage caps, since OpenAI responses aren't cheap at scale.

This project was made possible by Resend's new Inbound feature, which allows you to listen for webhook events triggered by emails sent to your domain. I've had a really good experience with Resend and was waiting for this feature to build out my idea.

I'm using GPT-5-mini to generate responses to emails, which is fast and affordable.

We'll see if this goes anywhere! It was a fun project to build.