Building software
for an agentic web.
We research and build small software, agent-native interfaces, MCP infrastructure, autonomous agents, and security systems for a web increasingly used by both humans and AI.
Areas of focus
Small SoftwareFocused software that solves narrow, practical problems without unnecessary complexity.WebMCPExploring how websites can expose capabilities directly to AI agents instead of forcing them to interact only through traditional user interfaces.MCPBuilding and experimenting with Model Context Protocol servers, tools, and integrations that let agents securely interact with software and data.AgentsResearching autonomous and semi-autonomous systems capable of executing real work across tools and environments.Agent SecurityExploring permissions, authorization, auditing, tool access, human oversight, and safer ways for agents to take actions.
Projects & Experiments
Our Thesis
Software is becoming dramatically cheaper to create.
As AI agents gain the ability to use tools, browse the web, call APIs, and execute workflows, software will increasingly be designed for both humans and machines.
We believe this creates opportunities for smaller software products, agent-native interfaces, new infrastructure, and entirely new approaches to software security.
Olive Research exists to explore that transition by building.