Dimension scores are derived from public data and fields; weighted into the composite. Reference only.
Launchboard is a company building products around “local-first AI.” The current page presents two product lines: the Boson AI mini PC and SiteSpawn AiCMS. Its core argument is that AI workloads should run on hardware users truly own, with data, models, context, and the software stack not dependent on cloud subscription services.
Boson is the hardware foundation. The page lists 126 TOPS of AI compute, an AMD Ryzen AI Max+ 395 processor, up to 128GB of LPDDR5x unified memory, and a built-in Bosonic k3s platform for self-hosted workloads. It also supports running models such as Qwen, Gemma, and DeepSeek locally via Ollama, with companion apps for Android, iOS, and Linux.
SiteSpawn is the software layer: an AI-native CMS with a CLI-first approach. Users can manage blogs, pages, stores, menus, media, forums, and other content from the command line. It supports JSON output, scripting, and version-controlled CSS, making it a better fit for developer workflows than for traditional visual website-building users.
Pricing information is limited. The page explicitly says “0¢ Subscriptions,” indicating opposition to subscription models, but it does not disclose the Boson hardware price, SiteSpawn licensing model, whether it is open source, or whether enterprise services or maintenance fees are available. Privacy is its strongest selling point: all computation runs on the user’s own hardware, data and context do not leave the network, and the product emphasizes no cloud dependency and no vendor lock-in.
The main advantage is its clear positioning: it combines local AI hardware, a self-hosting platform, and an AI CMS, making it suitable for developers who care about data sovereignty, controllable deployment, and local model capabilities. Design choices such as CLI, JSON, k3s, and Ollama also align well with engineering-oriented use cases.
The downside is that the publicly available information still feels more conceptual and specification-focused than practical. The page does not provide model inference speeds, content generation quality comparisons, deployment tutorials, after-sales support details, shipping coverage, or real-world case studies. Boson also mentions that a prototype has arrived, suggesting the product may still be at an early stage.
It is better suited to developers, homelab users, small teams building private AI applications, and people who want to manage website content using local models. Regular non-technical users may find the CLI-first approach too demanding. The page does not make it possible to assess accessibility from China, and payment methods or purchase channels are not disclosed. Alternatives include self-hosting Ollama/Open WebUI, using a local mini PC, or choosing self-hosted CMS options such as WordPress, Ghost, and Strapi.
⚠ This review is compiled from public sources and does not constitute a purchase recommendation. Verify all facts on the vendor's official site. Verify on launchboard.com official site.
launchboard.com is an Unknown Hardware & IoT provider. TG4G tracks its product information, an overall rating of 7.0/10, and a China-accessibility score of Workable. Click "Visit Official Site" to reach launchboard.com directly.