Avgidea brings private AI workflows to Apple Silicon Macs
Greenative Studio launched in Australia on 30 July 2026, giving businesses a local AI app for macOS that runs workflows on Apple Silicon Macs.
Avgidea, Inc., a Japan-based developer of data and AI platforms, built the desktop software for teams that want sensitive business data to stay off cloud AI services.
By default, prompts, responses, local files and fine-tuning data stay on the user’s Mac. Data only leaves the device if a team sets up external tools, APIs, databases or MCP servers.
That local-first design targets Australian organisations weighing data sovereignty issues and reforms to Australia’s privacy regime.
Users can download and manage open models from a desktop console, including Gemma, Phi and Mistral-based models.
Apple Silicon Mac workflows
The app includes local file search and summarisation, PDF-to-PPTX conversion, natural language database processing and local fine-tuning with a company’s own data.
Avgidea focuses on curated models released under permissive open-source licences, including Apache Licence 2.0 and the MIT Licence. As a result, businesses get clearer expectations for commercial use.
Founder Teppei Yagihashi said not every AI workflow belongs in the cloud. He argued that business documents, client files and internal data need an option that stays on a company’s own machine.
Greenative Studio costs US$18 a month or US$180 a year. It is available now through the Avgidea website with a free trial and no per-token usage charges.
For broader connections, the desktop app can link to Greenative Platform through greenative.mcp. Avgidea says that service can provide access to databases, cloud storage and internal systems.
Meanwhile, Avgidea offers Greenative Platform as a SaaS product or as dedicated customer-specific deployments. That gives businesses a local desktop setup first while still allowing extra integrations when needed.
Avgidea announced the launch from Tokyo, Japan, on 30 July 2026. The company described Greenative Studio as a private AI workbench with stronger control over where prompts, files, model outputs and fine-tuning data are processed.
