the infrastructure company of San Francisco
San Francisco, CA
Joined October 2020
- Tweets313
- Following90
- Followers490
- Likes261
opencomputer retweeted
Introducing PRumble - Bumble for PR reviews
- An @opencomputerHQ serverless agent reviews each PR at its head SHA in an isolated computer (clones the repo, traces callers, reads CI logs, runs tests)
- The verdict + red flags come back as typed JSON in a durable session
- A swipe runs gh pr merge with the GitHub App installation token
prumble.dev
opencomputer retweeted
Another day, another agent template. Do coding agents pick you?
1. Paste your URL
2. A bunch of prompts are generated and a fleet of agents run the prompts with different models
3. You can see who it installs, and who it considered.
Each agent is a few lines of TypeScript on @opencomputerHQ Serverless Agents, with credentials and streaming built in.
agentpicks.io
opencomputer retweeted
Meet Kevin - the Product Development Agent
when you truly care about the result, you don't just tell Claude "build it" and switch to something else, do you?
Kevin builds on this intuition.
github.com/diggerhq/opencomp…
opencomputer retweeted
Build your own software factory on @opencomputerHQ.
1. Cloud computers to clone, code, test and open PRs.
2. Durable sessions that keep context for follow-ups.
3. Built-in Slack, Linear, GitHub, cron and webhooks.
Try it today - opencomputer.dev/softwarefac…
opencomputer retweeted
A lot of GTM engineers are using cloud agents to improve AEO!
So I built PromptScout - it's @openseo running in an opencomputer.dev serverless agent!
1. Enter your website.
2. Find queries your competitors rank for.
3. Get content briefs, then deploy your own agent for weekly runs!
Try your free run and then run your own on a schedule by deploying the template!
promptscout.dev
opencomputer retweeted
Starting a series on how you can achieve the building blocks for this software factory on @opencomputerHQ serverless agents
Opencomputer gives you the building blocks to create your own factory with maximum flexibility:
- Memory
- Durable sessions
- Database
- Schedules + triggers
We will create a bunch of agents iteratively and build on top of our template until we have the full picture achieved. The first part of the template is the coder agent, which leverages opencode to perform pull requests.
Next up we will add memory to our coder agent. From there we will see how to include a
verification agent so that the PRs are also verified in a real computer for correctness (visual, functionality)
We will then create the PR review agent which checks the work of the coding agent
Next will be the oncall agent and the performance agent, two agents that check production metrics and react to incedents and suggest improvement performances
Bonus agent: the dreamer agent, which can run nightly, reflect on the memories and clean them up. It also has access to all the other agents' codebases so it can suggest PRs to improve itself and other agents!
Here's what OpenAI's agentic software factory looks like, today.
Details: newsletter.pragmaticengineer…
(thanks to all the OpenAI folks who explained how it works! And Perf Factory looks especially interesting to me)
opencomputer retweeted
in the inevitable near future all agents will go serverless
of course they need full linux computers! but ad-hoc not 24/7
and the probability of the harness getting access to its own environment must be precisely 0
we've built this
try now -> docs.opencomputer.dev
opencomputer retweeted
@opencomputerHQ is "heroku for agents"
- Write a typescript function
- Pick a model, add tools, skills/MCP.
- Deploy. Get managed execution, durable sessions and streaming.
Here's a demo of a template yesterday. URL --> Slide deck
Try it at shipslides.dev
opencomputer retweeted
was curious about how fast we can burst create @opencomputerHQ serverless agents so I did this experiment. Hitting 4000/s sessions created at peak
Each session is a full opencode session starting successfully.
Create your agent swarm on Opencomputer in a few lines of code .. leave the scaling and sandboxes to us
opencomputer retweeted
Gamma/Claude design style slides, but model agnostic and open source.
ShipSlides is a deployable template built on an @opencomputerHQ serverless agents.
URL or prompt → editable slides + PDF.
shipslides.dev
opencomputer retweeted
build background coding agents as easily as react
useTool() hook allows to enable tools conditionally
agents on opencomputer.dev - so much better
opencomputer retweeted
opencomputer.dev, September 2026
> best month yet on usage and revenue
> external agent sessions up 36x since launch week
> 28x in just the last two weeks
> most fun I've personally had building anything, ever
pumped for October. this feels like a real inflection point.
lfg
opencomputer retweeted
I built an @opencomputerHQ serverless agent template that tests your devtools quickstart on a schedule!
1. Reads your docs, then runs every step from scratch in its own isolated cloud computer
2. Runs daily and emails you only when it hits a new issue or blocker
3. Finds broken links, missing prereqs and auth walls before your users bounce
Try it on your docs (no login needed): axfeedback.dev
Here's a report it generated for @PlanetScale!
opencomputer retweeted
Dots was just launched at dev day in SF!
We made a model agnostic version you own, built on top of opencomputer.dev serverless agents
OpenDots: give it a task, it works on its own computer even after you close the tab, and reports back.
github.com/diggerhq/opend…
opencomputer retweeted
I genuinely still cannot get over how good opus 5.5 is at explainer videos.
Here's opencomputer.dev serverless agents in 60 seconds
(one shot opus 5.5 + eleven labs btw)
opencomputer retweeted
OpenComputer is open as in check our GitHub
but also open as in take the side panel off and OH WOW ALL THE NUTS AND BOLTS
not literally hardware ofc
rather, specialised cloud primitives
anyways
thats how your agents experience opencomputer.dev
tell them
opencomputer retweeted
Yes! You should try opencomputer.dev :)
1. Deploy agents like serverless functions. Write agent.ts, start sessions via SDK or HTTP.
2. Each session gets its own Linux VM.
3. Sessions are durable so stream/steer mid run + resume after hibernation.
4. API keys never enter the VM. Our egress proxy injects them after requests leave.
Users use it like a model agnostic alternative to Claude managed agents, primarily. Would love your feedback on it :)
opencomputer retweeted
a better way of building agents exists
it's Serverless Agents on opencomputer.dev
serverless because you aren't going to give every user of your product a beefy VM 24/7 - unless you are Meta
but your agent still needs access to files, git, etc
we solved this, try it
opencomputer retweeted
something something sell shovels when there's a gold rush!
- 13x signup growth this week
- thanks to a 400 point front page HN feature
- 3x growth in agent sessions created by users
lfg
opencomputer.dev
