Runs on a private AI box you own
One compact GB10-class device — 128 GB unified memory — runs it on-premise, sold and provisioned by us. No cloud API keys, no data leaving the building.


Run self-hosted AI apps on a private AI box
Most self-hostable AI tools are bring-your-own-model: they talk to any OpenAI-compatible endpoint. That means you can point them at a BrainOutput sovereign box on your own premises instead of a metered cloud API — same app, your hardware, your data.
Each page below explains what the tool is, its real license, how it connects to your box, and where BrainOutput fits. We sell and provision the box; the tools stay free and yours.
Chat interfaces
Coding agents
OpenCode
MITTerminal-native AI coding agent. MIT. Ships on our devices.
Aider
Apache-2.0Git-aware terminal pair-programmer. Apache-2.0.
Cline
Apache-2.0Autonomous VS Code coding agent. Apache-2.0.
Continue
Apache-2.0Open IDE assistant (VS Code / JetBrains). Apache-2.0.
Goose
Apache-2.0Extensible CLI/desktop AI agent. Apache-2.0.
Workflow automation
Agent platforms
Paperclip
MITOrchestrate teams of AI agents. MIT.
NemoClaw
Apache-2.0NVIDIA reference stack for sandboxed agents. Apache-2.0.
Dify
source-available (not OSI)Visual LLMOps for agents + RAG. Source-available.
Flowise
source-available (not OSI)Drag-and-drop LLM/agent flow builder. Source-available.
Langflow
MITVisual builder for agents + RAG. MIT.