I agree with all that Joseph commented above. I'd add that we're not removing users from the LinkedIn or Instagram scroll - if you want to view content on LinkedIn or any other social platform, you still need to go to the platform itself. We just wanted to make it easier to stay on top of the 5+ different messaging platforms we're all expected to check.
I think about it like an inverted WULPH.com (the office). I just want to have 1 app that is the source of all my comms.
We have been building Righthand for the past 2 years, and over the last 6 months, a lot of our conversations with customers have been happening over our personal phone numbers.
Like everyone else, we have a value to create obsessed customers but it's really hard to do that when you forget to text them because the thread is buried in your iMessage app.
We also use Superhuman, a thoughtfully crafted email client that actually helps you be "inbox zero," and thought to ourselves, what if we made an email client for iMessage? Thus, Attention was born.
Attention is a completely free iMessage client - we ask that you download it and give us feedback. It's been super helpful for us to stay on top of our own work comms, and we hope it does the same for you!
Privacy / Security: Your data is yours, fully local in LanceDB + SQLite, never leaving your machine unless you turn on the AI agent, at which point data just goes to whichever AI company you chose. We don't even have servers setup for the mirrors of our eventual cloud product yet (see below).
Free vs. Paid: Attention will always be free to run locally on your Mac with your own AI API key. If folks like the current app, we plan to add full support for other integrations (Gmail, Outlook, LinkedIn, Slack), an iPhone app, a wiki (like an Obsidian brain / PromptQL wiki) and other teams-specific features.
Existing Features: iMessage client. One-click MCP for iMessage. BYO Google Integration.
Request: please try it out and let us know what you think! Email us at founders at humans.righthand.ai.
TLDR:
- Problem: Work happens in iMessage but iMessage is a horrible client for it.
- Solution: Make a Superhuman-like client for iMessage.
We spent a year building guardrails around AI.
Then we spent the next year tearing them down.
After 2 years building virtual humans, here are our 3 biggest mistakes:
1. We underestimated how fast intelligence scales.
We spent months building scaffolding so our agents couldn’t break things.
A full week just to stop one from deleting a user’s inbox.
Then intelligence started doubling every ~6 months.
The guardrails became the bottleneck.
Now we spend more time removing constraints than adding them.
2. We treated memory like an engineering problem.
Vector databases solved recall.
But human memory isn’t just input → output.
It’s semantic search and graph traversal.
Backwards and forwards.
We tried every chunking strategy and embedding model.
Nothing worked until we rebuilt memory to mirror how humans actually think.
3. We tried to encode intent into workflows.
Zapier-style automation looks clean and reliable.
Until it isn’t.
A “simple” daily brief became 6+ nodes.
One failure downstream and I got a confident company summary… about a Montana farm.
Workflows don’t understand why they’re doing something.
So they can’t catch their own mistakes.
The lesson:
Treat the model as the system — not a component inside it.
I think about it like an inverted WULPH.com (the office). I just want to have 1 app that is the source of all my comms.
- Julius