.png)
Terminal Use recently launched!
TL;DR: Agent apps do not win on the model - they win on having a great harness. Terminal Use is agent hosting infrastructure that gives your coding agents the primitives they need to improve your agents.
👉 Ask: Know someone deciding what infrastructure to use for their background agent? Introduce them to the founders via email here or or book a demo here.
Founded by Vivek Raja, Stavros Filosidis & Filip Balucha
Hey everyone! Meet Vivek, Stavros and Filip. They’re excited to officially launch Terminal Use!
Vivek and Filip went to University together, and became close friends when they worked together at Palantir.
Vivek led the technical delivery of a large agent use case across US hospitals at Palantir. Stavros worked on infrastructure for Palantir dev tooling. Filip worked on the frontend for Foundry’s most-used app.
Over the past year, the founders built and deployed multiple agents following the same loop: build, eval, deploy, A/B test, generate evals from prod. Two things made this painful.
First, the tooling was fragmented: you had to stitch together multiple point solutions to cover the full lifecycle.
Second, production-grade orchestration for agents that used filesystems was hard. When they tried the Claude Agent SDK for real workloads, they hit issues: memory usage varies wildly with subagent count (risking OOMs), crashes require retry logic for durability, and all state lives in ~/.claude and the local filesystem, meaning you have to manage storing the filesystem to make use of it downstream.
At the same time, the founders all got a new powerful co-worker: Claude Code. They became curious about how they could create an out-of-the-box solution for their pains, and also leverage their new co-worker to iterate and improve our agents. This led to Terminal Use.
Terminal Use is infrastructure for background agents.
They provide a secure execution environment for your agent. They are framework agnostic - host agents built on Claude Agent SDK, Codex SDK, or your own framework. Bring your own dependencies too!
Terminal Use handles messaging for you - their messaging SDK lets you persist and stream messages back to your UIs. They also have a Vercel AI SDK integration.
To trigger your agent, you can use their SDKs, or run it on a schedule.
They’re purpose-built for agents that use filesystems. You can share one filesystem across multiple agents, and even fork filesystems to test different agents in parallel on the same input.
Finally, they know the pains of deploying agents to enterprises from their past experience, and they wanted to make it easy for you to develop agents for their businesses. Terminal Use allows you to physically isolate data and compute between your customers, and to permission resources granularly.
Sign up today at terminaluse.com. Or book a demo!
Internally, they use Terminal Use to ship Terminal Use! The founders believe in getting themselves out of the coding loop so that they’re able to ship faster. But vanilla coding agents keep you in the loop. They found that multi-agent configs, such as actor-critique loops lead to higher quality code and reduce the need for reviews.
Thus, they build an internal agent that spins up different coding harnesses that tjhey request for on the fly. For example, they ask their agent to spin up a Claude Code harness with Ralph Wiggum Loop. They’re then able to implement a plan with multiple different agents and choose the best implementation.
The founders love it so much that they’re soon to release their internal agent as a product.
Visit www.terminaluse.com to learn more.
Lorem ipsum dolor sit amet, consectetur adipiscing elit. Suspendisse varius enim in eros elementum tristique. Duis cursus, mi quis viverra ornare, eros dolor interdum nulla, ut commodo diam libero vitae erat. Aenean faucibus nibh et justo cursus id rutrum lorem imperdiet. Nunc ut sem vitae risus tristique posuere.
Built for Delaware C-Corps. Trusted by 1,000+ startups.
Copyright © Fondo (BloomJoy, Inc.)
.png)
