Ask HN: Who is using OpenClaw?

Well i built an equivalent of OpenClaw using Claude Code and hooking it up with WhatsApp. For mew I'm currently using it for following things

1. Morning brief + meeting preps 2. Managing client work and action items (tracking status, deliverables, etc) 3. Executing our AI workflows on my laptop. We have built several AI workflows for our agency and this setup gives the ability to seamlessly execute and control them through both mobile and desktop

Next on my to-do list is to build additional workfows for me and my wife around family logistics (travel, childcare, etc)

I tried it for a few days and then dropped it. The prompting/memory/context system for OpenClaw chewed tokens and performed subpar for my usecase. I was mostly interested in having AI writing little toy projects. I used pi over ssh instead, and eventually built myself a little mobile-first web UI wrapping pi.

I'm collecting caught-in-the wild use cases at https://www.clawdrop.org

I use it personally for cold outreach - specifically list building, enriching, and qualifying.

Yes - I've set it up as an 'office manager,' where it mainly snakily interacts with the local team via Slack, and controls an office TV to show our quote board, PTO calendar, and upcoming events. The Clawe is overkill for the use case, but sometimes is fun.

I've got a free hosted one from Zo Computer which sends me a daily news summary but I can't say I really use it. https://zo.computer/ if you want to try. They had $100 free credits for the promo code 'clawconlon'. Not sure if that's still going.

Their talk was quite good https://youtu.be/6rSEOzWY08U?t=2246

I used it, found it buggy, and quickly realized I could achieve everything it did with a long running Claude Code instance and a good mobile frontend. The joy is that you can customize everything to your hearts content just by asking Claude to build things for you.

- Background jobs? Cron? Huey + SQLite

- Memory? Create a job to write daily summaries to a memory/ folder

- Conversation log? Use hooks to write conversations to an sqlite file with full text search enabled

- SOUL.md? IDENTITY.md? USER.md? Stop wasting tokens and just use CLAUDE.md.

I only haven't quite figured out how to get channels working with 3rd party frontends.

I use claude code everyday. Most of my friend circle have a CC max subscription and we talk about and use AI all the time. Not a single one has installed openclaw yet.

For me personally I don't see that it can do a lot of things that CC/codex doesn't do and that _I_ want to do. Also I'm concerned about security.

For a while I wanted some agent I could tell what to do in my PC at home from my phone, so I just vibe coded a web site that can start CC and I used tailscale to secure it.

At one point, it was awesome.

then it became bloated.

then I both found and made gateway agent for different tools that does exactly one thing - connect me to the machine and does nothing else.

No, no subagents, no workflows, nothing, just be my gateway agent and the machine itself runs a whole other ecosystem. It is a nice wrapper/replacement on top of SSH, but only that. I don't need an ounce more.

I'm trying Hermes right now. I can't really find a good use for it. I tried to use it for research type stuff but Google Scholar literally by itself is faster, better, doesn't get rate-limited. Idk. I am pondering connecting it to my other bot I built that has more useful things like access to my thermal receipt printer and task management stuff but even that is kinda dumb because it already does everything I need, so I don't really know where I'm going with it. Honestly, I don't get the idea behind openclaw and hermes.

Nah, I deleted VM with it. My friends did it too. There is no real usage for it.

I have an instance that does search related to my research interests, tracks news related to viruses in the US, events happening around my area and an “urgent” news job that uses searx and for things going on around me. I used Qwen 3.5 9B and tuned some of the jobs with GPT 5.4. I recently switched to use Gemma 4 and there was seemingly no major difference. I’ve found it useful for the digest and for findings papers without much effort.

I use it. I think everyone should try the original version out before it gets too muddled so they get the spirit of the thing even if it doesn't end up being the thing -- in a safe space where you know what data you're allowing it to claw thru.

I wanted to try it recently and went to their doc website. Ar first glance it looked surprisingly detailed for such a young project but why not.

I didn't wanted to run a random, possibility vibe coded tool on my computer right away so I decided to run it in a container. That's an option, great ! Tu run it with docker they recommended an install script.

Out of habit I downloaded the script and started skimming through it to figure roughly what it would do to my machine.

The stuff was a mess. Most of it seemed overcomplicated code to do simple thing. Then I ran into inline python code in the shell script. That's not super elegant but there are some valid use cases, why not. Then there was inline JS code... WTF? Then we're back in inline python.

Seriously, that's just a script to run a docker compose file ! We're not yet at setting up providers or anything.

I was not to run that script on my computer, and was happy I decided to go the container way.

I got back to the website to see if they have manual instructions... They did, maybe there's hope after all ? Those instructions ended up being incomplete or broken, even for someone who does devops-y stuff on a daily basis.

I decided to drop the experiment there and it left me with the impression of a pile of code thrown together by an army of monkeys.

For me the prerequisite for leveraging openclaw was developing function oriented repositories of markdown files tied to my roles that capture pretty much everything I know about the subject and ongoing work, and working with agents as assistants off of those as context. As a founder, product manager, for growth etc.

From there it’s pretty natural that I wanted to talk to an always on agent not tied to any particular machine which has the same context plus access to google drive etc.

On a couple of different Ghost sites I maintain, I got suspicious sign-ups from travel-related domains that require email verification.

I suspect they might have used OpenClaw to accomplish the email verification step (no evidence) and were setting themselves up for later comment spam.

Even if you aren't using OpenClaw, it may be using you!

In Japan, it's a product targeted by the Ministry of Information and Communications, and some people without programming knowledge are touching it because it claims to automate things, but for now, I don't want to touch it.

I've mostly been using it as a calendar interface and email drafter. I also find it useful for a daily agenda with meals, meeting, weather and so on. I find the responds slowly (10-15s) so I've probably configured it poorly. I'm just using a Hetzner node and OpenRouter but maybe local is better?

I still use it.

We have it as a data analyst that’s been trained via exemplar queries and MD about the underlying databases.

It also does ad creatives analysis and overall paid marketing spend analysis, including delivering briefs and conversing with stakeholders about things such as cross-ad-platform deduplicated cost of customer acquisition.

In general it is used to lower the technical bar required for cross-domain collab.

there is a huge push in china and (tencent). but the main purpose is for the token market, to bring more users as a consumer of tokens. but i dont see any useful outcome from this yet.

tencent is anxiety about it's progress about AI, as Alibaba and ByteDance is iterated much faster. so before it has strcit rules for QQ/Wechat( has 1 billon + users) for AI bot , but now it's open to openclaw (some tencent version i guess).

but it's still a toy for most people.

How do y'all manage conversations with OpenClaw through platforms like WhatsApp? I mean conversation history and multiple conversations in parallel.

I gave up on openclaw but I’m presently installing Hermes.

I liked the idea, had it doing a few novel personal things, but it was so fragile and unknowable and 15% broken at every moment. It was expensive to keep and run, but I will essentially be running Hermes for free, so I’m cautiously optimistic.

I tried it a bit, and while the potential is huge, I mostly just use a cli agent (claude/codex) via Blink shell (iPhone/iPad) with Wireguard for technical work or https://agency.nu for any automations using integration, voice chat etc.

I used openclaw until claude released sheduled routines.

Nowadays i just create a repo insert context and then run sheduled routines with claude windows app against it.

For my use cases thats all i need and the most important part is that I can officially use my claude subscription instead of an API key.

It was very useful as an indication for who you can start ignoring in the ai hype space.

I'm only half joking. Blocked everyone / everything that hyped up openclaw and have been able to find much more interesting and reasonable ai related discussions in my feeds.

The people who are using it can't post here because OpenClaw deleted their HN account.

I see the hype, but for me I don't see the value yet. It seems neat, but what exact automation brought you real benefits?

I've used I cannot figure out the real benefit of it beyond novelty purposes.

I find Chrome Claude extension more useful for automating tasks online. Before ai I was writing my own macros which basically did the same thing in a more reliable deterministic way.

I tried using it for a specific web search task. I wrote a skill, got it all set up and deployed. It worked. But also, would have worked just as well as a cron job with some LLM looking at Brave API results. Like a lot of AI tools, it was a lot of work for underwhelming results.

Tried it in the earlier days and it performed badly. I didn't give it free reign on my computer due to obvious security concerns so sandboxed it to a docker container instead. I think for a lot of tasks it's probably more trouble to set this up than to just DIY it.

Telegram -> 1 Group/agent -> OpenClaw running on an old laptop

Using it for journaling and capturing ideas. Previous workflow as iPhone Memos. Now it's

voice message -> openclaw -> transcribe using parakeet -> git repo

I've tried it but after 10 minutes I found myself not comfortable running software this unpolished, specifically managing any sort of private data.

Tried it for a few weeks, was really buggy and unpredictable.

Kudos for the concept though, I ended up rolling my own agentic system with Claude Code from scratch that works much more reliably for my use cases.

On iPhone I use ChatGPT via Shortcuts and a-Shell for tool execution and Files for memory and state. I can schedule it to run or can invoke it from the home-screen via a shortcut.

This is a good example: https://www.lennysnewsletter.com/p/how-openclaw-changed-my-l...

I'm trying to. Currently there's a bug in the code that strips headers and doesn't allow me to authenticate to my AI Gateway service.

The whole thing is incredibly buggy. The dashboard is horrible, with page after page with similar-looking settings and what feels like hundreds of things I will never use. The categories in the dashbaord are also unintuitive. It's the kind of thing an AI would put together if it got very vague instructions. It doesn't scream quality and thoughtfullness. Not a bit.

IronClaw is much more promising imo. Trying it out right now. Much less issues so far.

I'm being forced to implement it at my job.

The hype is real. The repository says 5k+ issues and 5k+ prs. There's no way any human being knows what's going on in the codebase. How people trust this to hit their APIs and hang onto their API keys is beyond me.

This is yet another straw on the camel's back; I'm building my industry exit strategy because of AI.

I still try to figure out how to use it to its full potential.

I mainly run it through github-copilot/claude-sonnet-4.6 using GitHub Copilot Pro + at 39$/month

Task management: My entire todo system runs through GitHub Issues. I just tell it things like "mark that done" or "add a task for X" in Slack and it handles the gh CLI calls. Sounds trivial but removing the friction of opening a browser actually changed how consistently I maintain my list.

Morning/EOD briefings: Cron jobs post a structured summary to Slack every morning and evening — calendar, open GitHub issues, important emails. It pushes a RSS feed of my tasks that I can view on a widget on my phone.

Server management agent: I have a different agent which acts as the server admin. It runs Jellyfin, a few *arr apps, AdGuard, mealie, etc. I don't touch config files or docker compose manually anymore. I just describe what I want changed. I have it run its own security audits frequently.

I also have a personnal coach agent which tracks my weight, my weekly exercices using gcal and creates meal plans which gets pushed to mealie so I can know what to buy for grocery and what to cook.

Literature reviews: I describe a research question and it runs a full pipeline — searches Semantic Scholar + Google Scholar, creates a Zotero collection with clean metadata, then tries to fetch PDFs through 9 different strategies (institutional repos, arXiv, Unpaywall, EZproxy with my university credentials, etc.). Gets about 60-65% PDF coverage automatically.

I have a personal shopper agent called Betty which role is to get out there and find deals about stuff I want to buy.

I also use it to run data pipelines for research project. It's instructed to use opencode with openai/gpt-5.4 for coding with beads and gastown.

I still have to figure out how to manage model switching efficiently. I'm not there yet.

It's the first AI setup that genuinely changed how I work rather than just being a fancy search engine.

Yes, at our company we are using it very extensively. I genuinely believe we're near the forefront of usage. We have multiple isolated OpenClaw instances serving as employee within Slack.

I am using it as one the agent that is automating LinkedIn outreach by running a bash script & using ai wherever it needs some decision like finding first name or what message to write, etc.

ITT: "I don't use it, but ..."

I tinkered with Hermes yesterday but it still seems like a solution to a problem I don't have as a programmer.

i've been engineering things for almost 30 years and getting it wired up to Discord was worse than a root canal. Slack seemed just a tad better but it still doesn't even work.

I feel like most of this can be done with the platform tools at this point or a tiny bit of wiring of your own without the mega-bloat to make something generalized for the whole world.

I still use it daily, mostly for managing information consumption. It reads my twitter feed and scans HN twice per day, and sends me a digest of the discussions on Discord.

The best part is that it reads the comments too, and sends me a quick blurb. For example, this is what it sent me earlier, commenting on [1]:

  TL;DR: A classic essay arguing that compiler construction isn't as hard as thick textbooks suggest, pointing to Jack Crenshaw's accessible "Let's Build a Compiler" series as the real starting point.

  The Vibe: HN agrees most CS textbooks are overcomplicated — developers sharing their own minimal compiler projects and alternative learning resources.

It runs in an isolated vm and doesn't have access to anything other than my X account, so I'm not too worried about prompt injection. I also don't have any skills installed other the ones I developed or carefully vetted.

[1] https://news.ycombinator.com/item?id=47776796

a friend is using it but it seems like it breaks a lot.

ive got the scheduled claude-code running a couple scripts to find what events are going on round town and what food is cheap at grocery stores, but how much am i looking at the results? not super often. its publishing to a discord channel rhat makes it real hard to read

i use it with self hosted model, but often it's failed at using tools

its my remote control for claude code

it whatsapps me when its done or needs input it can not resolve, i start new session which are then done when i come to my computer

the reasons why i not use it more is tokens costs

yeah i could use a cheaper model for openclaw but then its just stupid

i am trying to run it on gamma next week

I have openclaw as a default install on all my dev servers. Pretty minimal setup with Telegram and Codex (since oauth is still supported). The setup comes in handy since openclaw can open and connect to tmux sessions and interact with them. I can pretty much do anything from telegram now.

I've had success using the underlying harness - pi-mono as a data analyst in a sandbox.