LongCut logo

Claude Code Outperforms OpenClaw When You Do These 5 Things

By Craig Hewitt

Summary

Topics Covered

  • Cloud VPS Enables Ubiquitous Claude Code Access
  • Claude Co-Work Schedules Tasks via Cron
  • Second Brain Repo Centralizes Life Context
  • Decision Journal Preserves Session Memory
  • Telegram Bot Feeds Mobile Notes to Claude Code

Full Transcript

OpenClaw is an amazing piece of software, but it can be frustrating, brittle, and really not very efficient as a work tool if you're not setting it

up just right. And for a lot of us, Claude Code is the answer for maximum productivity per hour spent. In this

video, I'm going to share five ways that I've kind of merged the concepts from OpenClaw that I really like into Clawed Code to make it more portable, make it

more accessible, and make it more userfriendly, especially when I'm on the go. And the first one here is giving me

go. And the first one here is giving me remote access to Claude code anywhere all the time. And so if you've set up your open claw on a VPS like a digital

ocean droplet or a hetner uh VPS, you know, it's always on, it's always there, it's always going, you lose power at your house, your Mac Mini that's sitting in the corner, doesn't really matter,

right? Uh I've done the same with Claude

right? Uh I've done the same with Claude Code, and I do most of my work in Claude Code in the cloud, and I just SSH from

my computer or from my phone. So I wrote about this on X and you can see the article here. or I'll link this in the

article here. or I'll link this in the description below, but basically I can SSH into a digital ocean droplet on my phone securely uh open a terminal app on

my phone and work in cloud code and it has all of the context, all of the setup, all the documents, everything exactly how I work at my computer just on my phone because I'm doing this in

the cloud. very much the setup of an

the cloud. very much the setup of an open claw where everything is living in one place with all of the context there and you're accessing it from just a different place. That's the gateway

different place. That's the gateway concept in openclaw. I'm basically

saying, hey, I like working in the terminal or an emulator. On my computer, I use Ghosty. On my phone, I use an app called Terminus, and it lets me go in

there. I have all the repos that I've

there. I have all the repos that I've cloned on this droplet. Uh, and so I'm working in the cloud most of the time, whether it's from my computer or from my

phone. So I have complete synchronicity

phone. So I have complete synchronicity across all of my Cloud Code projects, no matter where I'm working. Cool. I'll

link this article from X in the description below. Uh, and it walks you

description below. Uh, and it walks you step by step through how to do it. Cost

about $10 a month and it really transforms the portability and durability of kind of all of your Cloud Code sessions. Okay. And the second one

Code sessions. Okay. And the second one isn't actually in Claude Co. at all, but it's in Claude Co work. And I think secretly, or maybe not secretly, the

anthropic team after they didn't acquire OpenClaw and hire Peter Steinberger, the creator of OpenClaw, uh, have been incrementally

moving towards an OpenClawesque experience in their suite of products.

and in claude co-work which if you haven't used it yet it's kind of the more accessible child or sibling of claude code. So it runs in the claude

claude code. So it runs in the claude app on your desktop. Runs a lot like clawed code in that it's local. It has

access to any context in a folder you give it access to. But the real cool one here is you can go and say schedule this thing to happen. So much like we all

think it's magic that open clock can just do things magically at 2 a.m. in

the morning and then send us a report.

All it really is is cron and so uh so anthropic has created this concept in claude co-work

as well. So you could come over to

as well. So you could come over to claude code and just say set up a cron to run this code review scan every

morning at 3:00 a.m. if the computer is on. If the computer is not on do it as

on. If the computer is not on do it as soon as I wake up in the morning, right?

And so if I were to hit enter right now, like this would just be set up. It would

set up cron on my computer, which cron for those of you who don't know is just a technical term for like a schedule on a computer for a task to be run. Uh and

so you could just come into your your cloud code and just say hey like just do this thing for me. Open claw is not magic. It's taking some of the really

magic. It's taking some of the really smart concepts from these primitives like cron uh and just applying them automatically and kind of demystifying some of the more technical things. But

we can do all these with cloud code especially if your cloud code is always running in a VPS like I run mine often.

Okay. And another place where the anthropic and claude code teams are making claude code seem a little more openclawesque and always accessible is remote control.

And so if you come into your cloud code terminal here and just say slash remote control you can set up. Okay. So now I can code

wherever I want with this a continuation of this session. And so this is a great way. is you can come over to your phone

way. is you can come over to your phone and start coding in the Claude code app, the Claude app on your phone to continue the session that you had going here.

Really cool way to take something that you're doing on your desktop, move it to your phone so you continue, don't lose context, you don't lose all that kind of momentum that you have. and look for the

anthropic and the claude code and claude code work teams to add things like this continuously throughout the rest of this year and into next year to where the parody between openclaw and I think

claude code will close a lot and they'll look really similar but you can do a lot of these things today and the next one is more of a mindset thing than anything

which is having a repo or a project for cloud code on your computer called second brain you can call it whatever you want but the idea here is if you

would have gone to chatbt or claude chat in the web to do a thing, you would come here and do this. And this is where all of the context about your life and your

personal life and the decisions you're making, the things you're struggling with will happen. Uh, so, hey, I'm considering changing jobs. Hey, I'm

considering acquiring a business. Hey, I

want to lose 10 pounds. Hey, here's my P&L for the last year. Where am I losing money? Hey, um I want to create five new

money? Hey, um I want to create five new YouTube videos this week. What should I write about? Hey, pull in all my YouTube

write about? Hey, pull in all my YouTube analytics and let's idulate on what's going on here. Uh and actually this is a place where Claude Code and Open Claw run in parallel for me because my

openclaw has access to my second brain repo. And so I push this at the end of

repo. And so I push this at the end of every day to GitHub and my OpenClaw has access to that. So it just is sometimes I want to work in clawed code on my computer, but I want my openclaw Janet

to have context of all of that stuff.

And so I'm kind of mirroring these two.

You can see how these two systems are really coming together very nicely. And

I think at some point we kind of have to decide like, hey, I'm going to put all of my eggs in one basket or the other so I'm not duplicating things. But really,

I have minimized the amount of work I need to do between these two. But the

real big uh kind of mindset thing here is the second brain concept preserves all of the important context that you have as an operator, as a business

person, as an individual contributor in your company, as just a high agency kind of person. And the way that we're doing

of person. And the way that we're doing that is a skill that I created. And you

can just tell your clawed code to do this. And it's called decision journal.

this. And it's called decision journal.

This is kind of like the memory function inside of OpenClaw, but it's really specific. So, if I run decision journal,

specific. So, if I run decision journal, hey, log this conversation we were just talking about, it'll take all of the current context from a thread that I

have from a session in cloud code and log that just as a markdown file in the context folder. And so what this does is

context folder. And so what this does is as opposed to context being lost when a session ends or my computer decides to restart itself and lose all of that

session, uh it's preserved forever. So

if I go back 3 months from now and say, "Hey, remember when I uploaded my company's P&L? I want to take a look at

company's P&L? I want to take a look at how we can save an extra $5,000 this year." Or, "Hey, I want to continue the

year." Or, "Hey, I want to continue the discussion we had around increasing my thought leadership on LinkedIn." Or

whatever it is, right? You're not losing context and memory really between clawed code sessions if you're writing some of these important things that you've decided to memory and you're doing this

via some kind of uh slash command or agent that I call decision journal.

Okay. And the last one we're going to build here on the fly cuz I just had this idea. I don't know if you guys ever

this idea. I don't know if you guys ever do this. I had this idea. I wake up like

do this. I had this idea. I wake up like the other night and I'm like, "Oh, this would be the coolest thing ever. I

should do this." and we're going to build it live here uh on the on the video. Okay, so the concept is on my

video. Okay, so the concept is on my phone I I use my phone more for OpenClaw than I do my computer. I use Telegram and I can send a message from Telegram and Janet, my open claw, just sees it

and does things and magically kind of things happen. I want to be able to do

things happen. I want to be able to do that to Claude code as well. Um again,

kind of the parody between the two. I

haven't really fully bought into one camp or the other and I want to be able to persist information in both places.

And so what I want to do, I'm just going to describe it to Cloud Code, and it's just going to build it right here on the fly. So I'll use Super Whisper. You'll

fly. So I'll use Super Whisper. You'll

see a little thing pop up to where I talk into the microphone, it transcribes it onto my computer, so I don't have to type. But I'm just going to describe

type. But I'm just going to describe what I want. I want to create a system to where when I talk into a certain

Telegram bot or type something into a Telegram bot from my phone, it gets sent

to a VPS to log in a markdown file. And

then every 15 minutes, this Clawed Code project called Second Brain that we're in right now will look at that folder

and pull that in as context to store persistently in the context area. The

goal for this is sometimes I'm on the go and I want to just send a quick message to Claude Code, but Claude Code is running in the cloud and I don't want a

VPS. I don't want to log into a VPS or

VPS. I don't want to log into a VPS or maybe it's just a quick note or something and I just want to share a concept. I want to be able to send a

concept. I want to be able to send a message in a telegram bot and it gets stored in a way to where you in this claude code session or this repo called

second brain can access it and remember that context later or ask me about it if I ask you to set a reminder. So help me architect what that could look like.

Cool. And so Claude obviously really smart. Uh actually interestingly as of

smart. Uh actually interestingly as of the recording this I find Opus 4.6 got enormously dumber since the weekend. So

we'll see how it does here. But this is the kind of concept where uh you should take ideas you have in your head about how you can make your AI systems work

better for you in the way that you work and just talk to it about this stuff.

And so I just explained a very basic thing to Claude and it was like cool.

Uh, okay. So, architecture is Telegram toVPS second brain. Telegram bot, right?

That this lives in the VPS. Great.

Writes to this inbox. Cool. I just had a call with Daniel, right? Okay. So,

that's cool. You know, video idea blah blah blah. Okay. Now, sync VPS to this

blah blah. Okay. Now, sync VPS to this repo. Two options. The first is our sync

repo. Two options. The first is our sync cron, probably the simplest, and the other is like a GitHub action. I can see either of these working. And then a new file called context inbox messages that

gets pulled. You could also run a /inbox

gets pulled. You could also run a /inbox command that reads it and summarize these options.

So, so we'll run a Python script on a VPS. The Telegram bot appends that file,

VPS. The Telegram bot appends that file, a cron job that R sync, so it's like uh downloads from that onto this repo every 15 minutes, writes into this file, and

then I can run the slash command. Pretty

cool, right? So, like we'll set this up and do this, but this is just a way where you're like, hey, I don't want to set up a VPS. I don't want to be coding on my phone. I want to code and I want

to use clawed code on my computer only but I want the portability of openclaw without using openclaw maybe because truth be told like it's kind of a pain

right it's absolutely that kind of concept is the future but today without getting into openclaw and your own machine and all the security that goes along with it you could set up what

we're talking about here in 5 minutes and have all of the portability and accessibility of openclaw inside clawed code to say, "Hey, I want to send a message to where it has this context. It

has this thing. If it might, if I'm waiting for my kid after soccer, I can send this message and then when I'm back at my computer, it might ping me and be like, "Hey, I just pulled this bit of context you just sent. What do you want

to do with this?" That's the vision. And

that's how we all should be thinking about using AI agents in our business to be more productive. And so, as you can see, there's a lot of similarities and differences between Clawed Code and

OpenClaw. And I'm spending a lot of time

OpenClaw. And I'm spending a lot of time trying to get OpenClaw working better and better. And check out this video

and better. And check out this video over here where I share the seven things I've learned recently to make OpenClaw actually efficient and effective for me at

Loading...

Loading video analysis...