Vespper launches MCP tool for AI agents editing Word documents

By Splitfeed EditorialLast updated Sep 28, 2026, 18:11 UTC

On September 28, 2025, Vespper, a startup from the Y Combinator F24 batch, publicly launched a Model Context Protocol tool designed to let AI agents edit Word documents. The company was founded by Dudu and Topaz, with Topaz previously serving as a senior software engineer at the security firm Snyk. Vespper claims its product is three times faster, two times cheaper, and more accurate than the closest competing alternative. The tool is powered by a fine-tuned model the team developed after spending roughly a year building AI document editing software for pharmaceutical clients.

The launch sits at the intersection of two fast-moving areas: the MCP ecosystem, which is a protocol that allows AI agents to interact with external tools and services, and enterprise document workflows, which remain dominated by Word format files in regulated industries like pharmaceuticals. The practical bet Vespper is making is that AI agents will increasingly need to manipulate existing document formats rather than produce output in new ones, and that a purpose-built, fine-tuned model can outperform general-purpose alternatives on that specific task. The performance and cost claims are self-reported and have not been independently verified, which means the comparison to the unnamed nearest competitor, described in the launch as boom, is an open question.

No reactions from Pro-AI, Anti-AI, or Middle Ground camps have been published yet. Pro-AI commentators would typically welcome a tool like this as evidence that AI is becoming genuinely useful in high-stakes professional workflows, pointing to the pharma background as a sign of real-world validation. Anti-AI voices would likely raise concerns about AI agents making changes to legal or regulatory documents without sufficient human oversight, particularly in a sector where document accuracy carries compliance implications. Middle Ground observers would probably focus on the verification gap, arguing that independent benchmarks are needed before the performance claims can be taken at face value.

The argument will sharpen as independent developers and enterprises begin testing the tool against their own workloads. Benchmark results from third parties, or a formal evaluation comparing Vespper against other MCP-compatible document tools, would provide the clearest signal about whether the three-times and two-times claims hold outside controlled conditions.