YC F24 startup Vespper has launched a Model Context Protocol (MCP) server designed to simplify how AI agents edit Microsoft Word (.docx) files. The solution allows agents to interact with documents using a high-fidelity HTML abstraction, bypassing the need for agents to manage the complex and verbose XML structures of the Office Open XML (OOXML) specification directly.
Product LaunchesVespperMCP server
Vespper Launches MCP Server for High-Fidelity .docx Editing via HTML Reconciliation
The core technology utilizes a specialized "reconciler" model, trained using Low-Rank Adaptation (LoRA) on approximately 16,000 tasks. This model receives an agent's intended HTML changes and predicts the corresponding OOXML updates. By offloading the handling of document mechanics—such as runs, numbering, and styles—to this dedicated model, agents can preserve their context window for higher-level tasks.
In internal benchmarks comparing Vespper against five other solutions using GPT 5.6 models, the Vespper MCP demonstrated superior performance. The results showed that Vespper is 2.7 to 3.5 times faster and 2.7 to 2.9 times cheaper than traditional DOCX skill implementations, while maintaining higher accuracy. Notably, Vespper maintained a stable pass rate on longer documents where other solutions experienced significant performance drops.
The service is designed for high-stakes domains like legal, finance, and healthcare, where preserving document structure and formatting is essential. Vespper has released Python and TypeScript SDKs to integrate this capability into agent-based workflows.
Sources
- Launch HN: Vespper (YC F24) – SOTA Docx MCP （Hacker News Frontpage, 2026-09-28）
- Harvey Blog