We're a small team in San Francisco building the open platform for company agents. What's that you asked? Glad you asked, let us enumerate what it means to us.
We believe that all companies will eventually run on an intelligence that integrates with everything and does work for you, maximizing and focusing the output of human creativity. We don't aim to replace humans — we aim to build novel, delightful, and simple tooling that makes every day feel like the best day at work.
|
|
| Pairing after hours, Golden Gate in the window | Most decisions still happen standing at someone's desk |
Well you've caught us early, so... we're still pretty heads down building. But if you'd like an early scoop, here's what we're working on:
- An open-source "framework" and SDK. We aim to be a portable and open standard for building company agents. Unlike model labs, we want teams to own their agent and to understand how it works and to be able to customize it. Thus, we aim to be the most integrated and inter-operable framework that individuals can use to grow from a simple bot all the way to a full company agent that powers all teams at your company.
- An application layer. Model vendors focus largely on the models (checks out... ) but they're actively moving up into the application layer, things like GitHub agents, Slack agents, etc. We aim to provide open code and simple abstractions that just work to meet everyone in the company where they work, whether that's a GitHub agent surface, a Slack agent, a web UI "harness," and ultimately any surface that supports integrated agents.
- An open gateway. We don't think teams should have to be locked into the best of the frontier today (it changes! It's expensive! Opus is crazy verbose, and so forth) and should be able to route to the best model for the job, which may not always be the frontier. We expose an open gateway and enable teams to bring their own keys to support whatever model recipe feels right to them.
- Infrastructure primitives. We build robust platform infrastructure required to enable agents to do work securely, safely, and robustly and these power all of the above layers: we are building the connective tissue of a real company agent with our scalable sandboxes, filesystem volumes, and hosted UI product (aka a SaaS cloud product) that just works to host all of the above for teams that prefer a managed solution.
We're a small team in the Presidio of five engineers and a total team size of ten. If any and hopefully all of this sounds interesting, consider joining us and seeing if we have any open roles that pique your interest here.



