Computer-use automation
Coasty takes natural-language instructions and acts on a real computer instead of relying only on fixed selectors or API endpoints.
Coasty is a computer-use agent that automates tasks on a real computer through screenshots, clicks, typing, and terminal actions. It offers a native desktop app, browser access, and developer interfaces for workflows that do not fit standard API-based automation.
Coasty is a computer-use agent that operates a real computer through screenshots, clicks, typing, and terminal actions. The site positions it for tasks that are hard to handle with fixed integrations, such as legacy desktop software, mainframe terminals, internal portals, and web apps without an API.
It is offered as both a web experience and a native desktop app for Mac and Windows, with developer-facing access through a computer-use API and MCP server. The source also says runs can be recorded, replayed, paused for approval, and executed with your own model keys or in Coasty’s cloud.
Coasty takes natural-language instructions and acts on a real computer instead of relying only on fixed selectors or API endpoints.
The site presents Coasty as a computer-use API, with screenshots in and actions out through a REST-style interface.
The desktop app includes browser automation, desktop automation, terminal access, file operations, and auto-updates in one package.
The product supports human-in-the-loop pauses so a person can approve sensitive steps before the agent continues.
The product records each run step by step with screenshots, actions, and outcomes so work can be replayed and audited later.
The pricing and homepage copy indicate support for running multiple agents in parallel, including swarm-style backlogs across machines.
Use Coasty to read invoices, extract totals, match purchase orders, and route exceptions for review when the work lives in inboxes, PDFs, and portals.
Apply a plain-English procedure to a repeatable office task when the steps are stable but the underlying tools do not offer a clean API.
Run a worker against desktop software, mainframe terminals, or internal tools that are better operated by sight and click than by connector setup.
Pause before sensitive actions, then resume after approval for steps such as submitting or approving payments over a threshold.
Run multiple agents in parallel to clear a queue or backlog faster when a task can be split across machines.
The source shows browser, desktop, and terminal automation in one native app, plus a web interface for controlling the computer from anywhere.
The pricing page lists Starter, Plus, and Unlimited plans, and the Unlimited plan is billed monthly with no credit limits. The source does not provide more detailed plan rules beyond the published plan names and limits.
The download page says the desktop app supports Windows 10 or later on 64-bit x86_64 machines with at least 4 GB RAM, and macOS 11 or later on Apple Silicon or Intel Macs with at least 4 GB RAM.
The source says you can bring your own model keys and keep screenshots inside your environment when using your cloud or theirs, but it does not provide a full security specification on the pages provided.
The source shows a human-in-the-loop flow where an agent can pause for approval on selected steps and then continue after review.
CreateOS Sandbox is an isolated compute environment for running code and agent workloads inside Firecracker micro-VMs. It is designed for workflows that need machine-level isolation, private networking between sandboxes, and programmatic control through SDK, CLI, or MCP.
ByteAsk is a terminal-first AI coding agent for C and C++ that edits repositories and verifies changes with the real compiler, debugger, sanitizers, and tests before showing a diff. It offers a free tier plus paid plans, with editor connectors and zero-retention handling described in the source.
hob is an independent workspace for coding agents that keeps agent sessions, terminals, history, and follow-up work organized around the tools and providers you already use. It is aimed at developers who want local control over routing, history, and workspace structure rather than a bundled model stack.
Manta AI is an autonomous web app testing tool for teams that want to map application behavior, catch regressions, and generate tests without writing scripts or maintaining selectors. It works from a URL and supports plain-English test flows, run results with screenshots, and scheduled or deployment-triggered checks.
Redline is a budgeting tool for Claude Code that paces sessions to stay within time, token, cost, or plan-percentage limits. It uses Claude Code’s native hooks and statusline to help sessions finish with a usable result instead of stopping abruptly.
SonOf connects to your repo and PM tool, audits the codebase and surrounding product context, and turns approved work into shipped tickets with senior engineering review. It is aimed at founders and engineering leaders who need backlog help without hiring a full team immediately.