GenOffice with LM Studio: local AI setup tutorial
Connect GenOffice to LM Studio with CORS enabled and a /v1 endpoint, then follow document, XLSX analysis and HTML generation examples.
Player not loading? Watch on YouTube
This tutorial connects GenOffice's integrated AI agent to LM Studio for document work with a local LLM. The presenter describes GenOffice as an open source office suite available for Windows, Mac and Linux. The examples cover writing a document, summarizing an XLSX spreadsheet with made-up data and generating a basic website for a solar panel business.
The setup uses quantized Gemma 4 models suited to tool calling. The supplied guide names jmingo/gemma-4-e4b-it-qat and instructs users to enable Developer Mode, disable Use LM Studio Engine Protocol and choose model load parameters manually. Context length and GPU offload depend on the available graphics hardware and RAM. The local server needs CORS enabled and its status set to Running.
In GenOffice, the guide selects Custom as the provider and uses http://127.0.0.1:1234/v1 as the base URL. It sets maximum output tokens to 35,000, disables Genspark cloud tools, then saves the model selection and tests the connection.
The presenter reports successful document generation and spreadsheet analysis, but describes the generated website as basic rather than top-tier work. Other API providers are also available. The speaker presents local processing as a way to keep documents private; the tutorial does not establish a universal privacy guarantee.