Practical overview
When it fits
Choose FutureOS when a single agent should span several interfaces and its access to files or shell tools needs configurable approval and sandbox boundaries. It is also suited to users who want to bring their own model provider or define a compatible custom endpoint.
Who it is for
- Developers building or adapting multi-interface AI agent experiences.
- Advanced users who want a local-first agent with configurable model providers.
- Teams or individuals working with repositories and tools that require explicit file or shell-access boundaries.
- Users who need to continue agent sessions across terminal, desktop, mobile or chat surfaces.
Documented uses
- Configure an agent with hosted, listed or custom model providers.
- Work with files, shell tools and other agent skills under approval or sandbox settings.
- Access shared sessions, memory and skills through multiple client surfaces.
- Run longer-horizon goals with durable state, monitors and verification gates.
- Fork and navigate conversations through JSONL session history.
Requirements & limits
The repository is public and open source. The README documents installers for macOS, Linux and Windows, with platform-specific desktop and CLI packaging details.
- The shared architecture is based on a Rust core and a gRPC agent serving multiple interfaces.
- Model access can use FutureOS hosted models, built-in provider entries, or custom OpenAI- and Anthropic-compatible endpoints.
- Tool boundaries are configurable, but sandbox mechanisms and guarantees vary by operating system.
- The loop control plane is designed for long-horizon runs with durable state and verification gates.
- Adding another client surface is Likely compatible with the architecture, but the supplied documentation does not show an interface-extension API.
Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.
Install FutureOS and configure a model
Set up the documented FutureOS installation and provide at least one model provider.
Before you start
- A supported macOS, Linux or Windows environment.
- A FutureOS provider account for hosted models, or credentials for a listed or custom provider.
- Internet access for the documented installer or provider setup.
Steps
Run the documented macOS/Linux installer or Windows PowerShell installer.
Complete the installer’s
future initstep.Run
future configto configure the model provider.For hosted FutureOS models, run
future auth login; alternatively configure provider credentials in~/.future/agent/auth.jsonor custom providers in~/.future/agent/models.json.Start using the configured agent through an available interface.
Expected outcome
FutureOS is installed and has at least one configured model provider for agent responses.
Guide limitations
- The supplied installation text documents setup but does not provide a complete interface-specific usage procedure.
- Sandbox availability and guarantees differ by operating system.
- The project documentation does not establish pricing or general service availability.
Sources & review
Source record details
- Source availability
- Active
- Sources checked
- 3 October 2026
- Original website title
- futuregene/future-os · Official homepage title · retained in its source language

