Developers

Deploy HTX301, from unbox to first token

Everything you need to run large models on your own hardware: quickstarts, an OpenAI-compatible SDK, a supported-models list, and a community of operators.

Read the docs Join the community

Documentation and downloads are rolling out alongside HTX301 early access. Items marked soon are in progress.

Documentation

Self-serve, start to finish

Everything to get your hardware running and integrated. Start here.

Quickstart

From unboxing to first token — install the card, drivers, and run your first model.

Supported models

The living list of models validated on HTX301, with performance notes.

SDK & API reference

Serve with an OpenAI-compatible API; integrate with Triton, LangChain, and your stack.

Deployment guides

soon

On-prem sizing, multi-card scaling, and tuning for throughput vs. latency.

Drivers & firmware

soon

Versioned downloads with release notes and changelogs.

Troubleshooting

soon

Common issues, diagnostics, and how to get help.

Open by default where it counts

SDK & examples on GitHub

The enablement layer — SDK, deployment examples, and the supported-models list — lives in the open so you can evaluate before you buy and integrate without gatekeeping. Fork it, file issues, send pull requests.

github.com/skymizer

GitHub organization coming online with HTX301 early access.

# OpenAI-compatible endpoint on your own HTX301
from openai import OpenAI

client = OpenAI(base_url="http://localhost:8000/v1",
                api_key="local")

resp = client.chat.completions.create(
    model="llama-3-70b",
    messages=[{"role": "user",
               "content": "Summarize on-prem."}],
)
print(resp.choices[0].message.content)

Community

Operators helping operators

We're building the community where it already lives, and growing our own as the deployed base grows.

Live at launch

Discord

Real-time help, early-access channels, and a direct line to the team. The fastest way to get unblocked.

Live at launch

GitHub Discussions

Async questions, feature requests, and show-and-tell — searchable and tied to the code.

Planned

Community forum

A dedicated Q&A forum will open once there are enough deployments for operators to answer each other.

Support

Where to get help

Technical & software

Docs, GitHub Issues, and Discord for setup, models, and the SDK.

Hardware & warranty

Handled through your point of purchase. Email us if you're not sure where to start.

Everything else

Email contact@skymizer.ai and we'll route you.