@OsaurusAI
Account based inUnited States
About this account
- Account based in
- United States
- Connected via
- Web
Account-level information from X, not a live location or the device used for a specific post.
Own your AI. Agents that remember, execute code in isolated VMs, and stay reachable from anywhere -- all on your Mac. Any model. No cloud required. Open source.
Joined May 2025
- Tweets1.4K
- Following30
- Followers10.7K
- Likes1.3K
Pinned Tweet
Models get commoditized.
The harness compounds.
Thanks @SarahPerezTC for the writeup.
Osaurus brings both local and cloud AI models to your Mac techcrunch.com/2026/05/15/os…
Most local AI demos run on maxed-out Macs.
40% of our users have 16GB or less.
Raptor is for them. 4B. Wi-Fi off.
Raptor 0.6, 4B, JANG_6M quant.
Reads xlsx and docx through native file tools. Around 40 tok/s.
Built by @dealignai.
osaurus.ai/models/raptor-06
Osaurus retweeted
#design #ux #ai #localai
I just had a great conversation with the founder of @OsaurusAI
I’ve been following the evolution of the app since the beginning of the year. They built Raptor models, and the results are impressive.
osaurus.ai
Osaurus retweeted
A little verbose for my taste, but absolutely gets the job done. Try out the ~6gb and ~8gb quants of Bonsai 2 27b on your Mac!
Bonsai 2 dropped hours ago. It's already doing real work in Osaurus.
Read two CSVs. Wrote the Python. Drew the chart.
Then found the one channel losing money and told me why.
27B. Ternary. Nothing left the Mac.
Today, we’re announcing Ternary Bonsai 2 27B.
Based on Qwen3.8 27B, Bonsai 2 27B is 9x smaller than its full-precision counterpart while retaining 98.2% of its aggregate benchmark performance.
Two months after the first Bonsai 27B release, the biggest change is quality. The footprint remains 5.9 GB, but the gap to full precision has narrowed materially, with particularly strong gains in agentic coding, multimodal reasoning, and long-horizon tool use.
Ternary Bonsai 2 27B is available today under Apache 2.0.
PrismML said the big Bonsai 2 gains were agentic coding and tool use.
This is what that looks like on a Mac.
The model wrote the code, ran it in an isolated sandbox, and handed back a real PNG. Not a code block. A file.
Weights, ready for Osaurus:
huggingface.co/OsaurusAI/Bon…
JANG conversion by @dealignai. Model by @PrismML.
Osaurus is free, open source, and runs this on your Mac today.
github.com/osaurus-ai/osauru…
Local AI shouldn't be limited to expensive hardware. Everyone should have the opportunity to own their AI.
Osaurus retweeted
This will be coming to Osaurus for Windows as well. Not everyone has an RTX 30/40/5090. It'll open up local model usage for PCs with more common GPUs. You won't need a workstation to use local AI.
Raptor 0.6 just shipped.
Most Macs have 8 or 16 GB. Most AI models pretend they don't. Raptor is our agent model built for those machines. Reads folders, checks spreadsheets, drafts emails. Local, free.
3.7 GB, down 41% from the last version.
What changed in Raptor 0.6:
New base: Spark-X2.5-4B, 4B dense, all parameters active
0.5 was 8B MoE with ~1B active, 6.3 GB
0.6 is 3.7 GB, 1M native context, ~105 tok/s on M5 Max
Rank-4 LoRA on attention only, tuned on the Osaurus tool surface
No repetition penalty needed. The loops are gone.
Crafted by @dealignai
osaurus.ai/models/raptor-06
Raptor runs inside Osaurus, the part you actually own.
Tools, memory, sandbox, approval gate, native Mac integrations. Open source, MIT, no account for local use.
Most Macs are 8 or 16 GB. We'll keep building for them.
github.com/osaurus-ai/osauru…
Osaurus retweeted
Windows users deserve a great AI app that doesn't hog their system resources. I built the @Windows version of Osaurus to be fully native with a focus on user experience, performance, and ease of use with local models.
I can't wait for you to try it and share your feedback!
Osaurus retweeted
Lets be honest, every other engine having so many flags to be able to even launch, then handling all the json configuration to connect to the harness so you can actually use it; who wants to really do all that?
This has to for sure be the cleanest and easiest way to be able to just directly launch an LLM and use for real scenarios, now available for Windows!
@SonofNun did some amazing work
Osaurus for Windows is built with WinUI & C# on .NET 10 and runs on x64 and ARM64. It idles ~120MB, leaving your RAM for the rest of the work you're doing.
Use a Local model, bring your own API key, or use your existing subscription.
Come help shape the future of local AI on Windows. Drop your email here for access to the beta: osaurus.ai/windows
The macOS version is open source, MIT licensed: github.com/osaurus-ai/osauru…
It's here. The first fully native agent harness, built for Windows. No electron in sight.
Osaurus for Windows is almost ready for beta. Get ready for a team of agents running right on your PC, local and private.
Watch Qwen3.8-27B set up a team of agents to perform a code review.
We've been building exactly this while being fully open source and MIT licensed.
In Osaurus, you can build a team of agents where each one has its own memory, tools, and skills. An orchestrator can delegate work to them as sandboxed subagents. You can tie them together with Projects to share the same instructions, knowledge base, and memory across every chat.
Our harness is fully model agnostic (local or bring your own cloud key)
github.com/osaurus-ai/osauru…
