[bigpapikite]

I don't personally know many people who've used it so I'm not sure if this was a me thing but here was my experience in short:

I set up OpenClaw on a raspberry pi 4 that I could ssh into using my main computer. My main goal for using OpenClaw was just as a morning debriefer that could scan my google calendar, trello board, and gmail to let me know what I had happening for the day and also weekly to give me a forecast for the weeks ahead to see how busy my month was. I spent about 40-50 bucks in one week just working through kinks and having it fix itself until I stumbled onto a post that helped me optimize my model usage for price instead of just throwing Opus and Sonnet at everything.

Even after making this adjustment, the morning debriefer worked maybe once or twice a week and broke every other morning, telling me that it fixed itself and it would never happen again. At a certain point I just got fed up with it and cut the cron job, it's still running on my pi but I never use it.

Pretty sure Claude has something like this now but I'm pretty thrown off the whole thing, I'd rather just take the 30-45mins to plan out my day in the morning myself.

[littlemishka]

I work on something called Town (town.com) and your morning debriefer is basically a preset workflow we ship out of the box. You connect your Google Calendar and email, flip it on, and it just sends you a daily rundown. No cron jobs, no token tuning, no "I fixed myself" lies. Took us a while to get the reliability right but it genuinely doesn't break every other morning. It's free still for early tester, feel free to give it a shot!

[jimmygrapes]

> No X, no Y, no Z.

Please, everyone, stop this.

[Biologist123]

Well done. This is precisely what the future will look like. I’ve hacked together my own version of this using N8N whereby I feed in tasks via text or photo through telegram and it outputs a to do list. At some point I’ll link it to my personal wiki and then have it action tasks. This was just a fun project, but one design decision is to keep third party services away from commercially sensitive emails.

[robryan]

I had a surprisingly large amount of issues getting it to correctly message me on discord if needed as part of heartbeat. Multiple times it broke and we fixed it. Then it was chewing through way too much usage so I set it to 6 hourly heartbeat and it stopped messaging entirely again, haven't bothered to fix.

It is nice though for debugging home server things when I ask it to.

[haneul]

Works perfectly installed as part of a server and at mentioning you there. Direct DMs are finicky in my experience.

[hamasho]

Do you just give your Google account to OpenClaw, or create a separate account with limited permission? I'm worried that OpenClaw decides to create an entire website on GCP project without asking if they see a message like "have you already developed and deployed the management dashboard?"

[cadamsdotcom]

Why not use Claude to write yourself a daily briefing, and deliver it to your email or someplace you’ll receive it?

[SyneRyder]

That's what I ended up doing, made my daily briefing a Cron job and running it via "claude -p". Wired it up to make a podcast, with MCP tools I made to create an MP3 with OpenAI, another to upload it to one of my sites with an updated RSS feed, so I can listen in the AntennaPod podcast app each morning.

[cadamsdotcom]

Nice. Even better would be having your agent write code for the deterministic bits and telling the agent it should “invoke the script called blah” to do uploads (or whatever you want to have happen deterministically).

[SyneRyder]

Yep, I agree! My MCP tools are local compiled Go binaries, and the tool that uploads my podcast is actually a local Go CLI that Claude calls. Claude's main role / intelligence is in evaluating which of the morning's HN & Lobsters news is most relevant to me specifically, and writing the podcast script. I'm all for deterministic tools, and it saves on tokens too.

One advantage of splitting it into MCP tools though, one day I'd run out of pre-paid OpenAI TTS credit, and Claude was smart enough to try using Mistral TTS instead. I could have done that fallback deterministically too, but it wasn't something I'd thought of yet.

I once had a friend tell me they'd got their AI to tell them the weather every morning... and the thought of that poor AI, web researching Weather APIs & writing a new python script to call the API every morning, instead of just doing the research once and making it a binary (or even just a curl line)... drove me crazy. All that wasted time and compute. Some people just like to watch tokens burn.

[jpfaraco]

Interesting. How's the performance on the Pi? I also have a Pi 4 running a few containers and thought about running Openclaw on it, but never got around to it because of performance concerns. I ended up installing Nullclaw which is simpler but much more lightweight.

[organsnyder]

I have a series of skills and scripts that I manually invoke using Pi. It checks my calendar, Slack, email, etc. and keeps pretty good track of my day, to the point that I trust it enough to get a lot of noise out of my head.

[7thpower]

What kind of noise does it filter out for you?

[organsnyder]

Mainly all the tasks that I otherwise feel I have to track myself. I've found that my head feeling clearer is an indication that I trust a tool to not lose context.

[godot]

I think this is a pretty common experience with setting up openclaw to do useful work. Both the unreliability as well as all the reassuring you it fixed some stuff but not really.

[swyx]

agree. it breaks a lot, integrations are hard, it turns out, and openclaw can only fix itself so much before you need human intervention and it just stops