// open source · BETA

AI shouldn't have a meter.

Unlimited tokens. Forever.

Your machine. Your agent. Use it from anywhere. OpenMonoAgent.ai is a terminal-native coding agent powered by local LLMs: 100% open source, free forever, and installed with a single command. Proudly built on C#/.NET, because AI tooling should be infrastructure, not a subscription.

Want it on your own servers? We install and support private OpenMonoAgent deployments.

Setup instructionsNeed Help?View sourceBETAMobile appBETA
A PROJECT BYStartupHakk
~/my-project - openmono
$bash <(curl -fsSL https://raw.githubusercontent.com/StartupHakk/OpenMonoAgent.ai/refs/heads/main/get-openmono.sh)✓ installed. # No API keys. No cloud. Just your machine.$openmono agent→ Detected Docker, NVIDIA runtime→ Starting llama-server :7474 ✓ Ready. 21 tools + MCP registered. Context: 196k. you ›refactor Program.cs to use async/await→ reading Program.cs (142 lines)→ analyzing with Roslyn (blast-radius: 3 call sites)→ proposing edit… [approve? y/N]✓ Patched. 0 bytes left your machine.
// STARTUPHAKK MANIFESTO

"AI shouldn't be a subscription you rent. It should be infrastructure you own - sitting on your desk, serving your code, answering only to you."

- The StartupHakk democratize-AI thesis

StartupHakk

We also install and support private OpenMonoAgent deployments on your own hardware.Get in touch.

  • 01
    Local-first, alwaysThe model runs on your hardware. The agent runs on your hardware. No one else sees your code.
  • 02
    Unlimited tokens - Zero costEvery prompt after the install is free. No per-token billing. No rate limits. No surprise invoices.
  • 03
    Sandboxed by defaultDocker-native. The agent mounts your project in and can't escape. Permission gates on every destructive op.
  • 04
    Open sourceRead the code. Fork the code. Run it offline forever. That's the deal. Powered by C#/.NET.
// benchmarks
HardwareMemoryModeltok/s
Ryzen 9 7940HS / 780m
DDR5 dual-channel		32 GB RAMTernary-Bonsai 2 27BFULL ACCURACY~8 to 10
Apple M5 Pro
307 GB/s bandwidth		64 GB RAMiQwen 3.6 35B A3BFULL ACCURACY~45 to 48
RTX 3090
936 GB/s bandwidth		24 GB VRAMQwen 3.8 27BFULL ACCURACY~42 to 45
RTX 3090 RECOMMENDED
936 GB/s bandwidth		24 GB VRAM
64 GB+ system RAM		Qwen 3.8 Flash-NextFULL ACCURACY~55 to 60
RTX 4090
1,008 GB/s bandwidth		24 GB VRAMQwen 3.8 27BFULL ACCURACY~47 to 50
RTX 5090
1,792 GB/s bandwidth		32 GB VRAM
64 GB+ system RAM		Qwen 3.8 Flash-NextFULL ACCURACY~180
RTX 3060
360 GB/s bandwidth		12 GB VRAMTernary-Bonsai 2 27BFULL ACCURACY~20 to 32
RTX 4060 Ti → 4080
288 to 717 GB/s bandwidth		16 GB VRAMQwen 3.8 27BFULL ACCURACY~20 to 42

Runs fast on modest hardware.

llama.cpp is built in. Setup reads your hardware and picks the best model for it. GPU runs Qwen 3.8 27B, or Qwen 3.8 Flash-Next when a 24 GB card has 48 GB+ of system RAM behind it. SoC machines like the Ryzen 9 7940HS run Bonsai 2 27B, where speed tracks RAM bandwidth. Smaller GPUs get the same 27B, quantized to fit their VRAM. No rate limits, no clock watching, no bills.

The sweet spot? A used RTX 3090. Around $2,000, it delivers 45 to 60 tok/s, indistinguishable from a cloud API. But you don't need a GPU to get started. An APU box like the Ryzen 9 7940HS with 32 GB RAM runs the lean Bonsai model at a steady 8 to 10 tok/s, perfectly usable, completely free, sitting on your desk. Pick your hardware. We handle the rest.

Not sure what hardware you need?

Tell us your budget and machine. We'll recommend the right model and setup.

Need Help?
// new · VS CODE EXTENSION

The full agent
experience,
in your editor.

OpenMono now runs inside VS Code and Cursor: local, private, no cloud. Agent chat, Plan/Build modes, permission management, file references with @, and sub-agent support. All the power of OpenMonoAgent.ai, right where you code.

Install ExtensionLearn more →Need Help?
OpenMono Plan Mode showing plan output with decision buttons in VS Code
// capabilities

A full coding agent
that lives in your box.

Everything you'd expect from a modern AI coding tool, and a few things you wouldn't, because it doesn't have to phone home. No complex setup required.

01
⚙
NEW

Host mode

Need the agent past the sandbox: services, installs, host logs, docker? Host mode lets it, but every machine-level action goes through a HostExec call you approve first, one at a time. Powerful when you want it, contained when you don't.

02
✦
NEW

Now running the Qwen 3.8 family

Our strongest models yet. Qwen 3.8 27B dense on GPU, Qwen 3.8 Flash-Next streaming from a 48 GB-class box, and Ternary-Bonsai 2 27B where VRAM is tight. Each a big jump over the last generation on agentic coding benchmarks. Auto-picked for your hardware, free to run.

03
⊞
NEW

VS Code Extension

Full agent chat, file references with @, Plan/Build modes, and permission management, directly in your editor. Runs locally, stays private.

04
⎇

Embedded inference, zero setup

Auto-detects your hardware and loads the right model for it. GPU runs Qwen 3.8 27B at full power. CPU drops to Qwen 3.6 35B A3B, lean and still capable. No switching, no config, no API keys. Just run it and it figures itself out. Fully offline.

05
⧗

TUI built for long sessions

Full-screen ANSI TUI with streaming tokens, live tok/s meter, context-window usage, auto-compaction, and a flipping hourglass that actually tells you the agent is still thinking.

06
◰

Docker-sandboxed

The agent runs in a container. Your project is bind-mounted in. It can edit your files - but it can't leave the box.

07
⚒

21 tools + MCP

File I/O, shell, search, web fetch, LSP, patches, sub-agents, plan mode. Extend via Model Context Protocol servers.

08
⌕

Built for .NET, focused on .NET

OpenMonoAgent isn't just built on .NET, it's designed for .NET developers first. Roslyn is baked in, giving the agent real compiler intelligence: type hierarchies, call graphs, cross-assembly references. The agent understands your solution the way the compiler does.

09
☷

LSP for C# and TypeScript

Hover, go-to-definition, and references for C# and TypeScript. Real code intelligence, locally.

10
▤

Playbooks

Typed, composable, stateful workflow automation. Step sequencing, gates, and templates - not just markdown recipes.

11
⎈

Dual-box mode

Run the agent on your laptop, inference on your home GPU rig. Outbound-only via relay - works behind any NAT or CGNAT. Signup at app.openmonoagent.ai

12
◷

Persistent sessions

JSONL transcripts, cross-session memory, auto-compaction, and snapshot-based file undo. Nothing lost, nothing leaked.

Need help using any of these features?

Tell us what you're building. We'll help you wire it up.

Need Help?
// the landscape

How it stacks up.

Great coding agents already exist. They're also cloud-locked, metered, and running in someone else's datacenter. We built the third option.

Claude CodeOpenCodeOpenMonoAgent.ai
CostPer-tokenPer-tokenUnlimited tokens - Free
PrivacyCloud-onlyCloud + OfflineFully offline (private)
Ease of InstallSimpleDifficultSimple
Local LLM EnvironmentNoneAgentAgent + Inference
Tools441521 + 22 via MCP
Written innpmnpmC#/.NET
SandboxingHost installHost installDocker-native
LicenseCommercialOpen sourceopen source

Still comparing options?

We're happy to walk you through the differences and what works for your setup.

Need Help?
// install

A FREE COURSE on how to get up and running with OMA, build your own site, and your own web-game, publish it live on the web.

Start from zero. We install OpenMonoAgent.ai on your machine - the model stays local, the agent stays local, your code never leaves your network. Then you use that stack to build and publish a real website and a real browser game to the live web.

Start the free course

Need help with the course setup?

We'll walk you through the install so you can start building day one.

Need Help?
Watch on X

built by

StartupHakk
┌─────────────────────────────────────────────┐
│    OPEN.    LOCAL.    YOURS.    FOREVER.    │
└─────────────────────────────────────────────┘
  
┌────────────┐
│   OPEN.    │
│   LOCAL.   │
│   YOURS.   │
│  FOREVER.  │
└────────────┘
STARTUPHAKK × OPENMONOAGENT.AI

Stop renting
intelligence.

Clone the repo. Run the installer. Own your AI dev stack before the end of the afternoon.

OpenMonoAgent agent flow diagram
Star on GitHubEnter giveawayNeed Help?

OpenMonoAgent
Now on mobile.

Direct your self-hosted inference server from anywhere. Full control over your local AI dev stack - no cloud required.

Download on the App StoreGet it on Google Play

Need help connecting the app to your server?

We'll help you point your phone at your own inference box.

Need Help?