Paperzilla

What Paperzilla does

Paperzilla connects AI assistants to focused feeds of research papers and their associated digests. Its primary use is research discovery: an agent can find public research digests, search feeds belonging to the user, and retrieve paper content in Markdown.

The service is oriented toward recurring, focused research rather than broad web search. Morning research briefs provide a way to review new material from selected paper feeds, while feed search helps an agent look for relevant content across the feeds available to the user.

How it works

Paperzilla exposes its research functions through the Model Context Protocol and uses OAuth for access. An MCP-compatible assistant can use the available Paperzilla capabilities as part of a conversation or research workflow, subject to the access granted during authentication.

The content model is built around paper feeds, public digests, and Markdown versions of papers. Public digests support discovery without requiring an agent to start from an individual paper. Searching a user's own feeds provides a more targeted path when the research source has already been selected. Reading paper Markdown gives the assistant text it can process directly.

Tools and capabilities

Paperzilla provides three central research capabilities:

• Discover public digests from focused paper feeds.
• Search the user's own paper feeds.
• Read papers in Markdown format.

These capabilities support complementary stages of a research task. Public digest discovery can surface available collections and recurring briefings. Feed search narrows the search to sources the user follows or maintains. Markdown access supplies paper text for reading and further analysis by the connected assistant.

Morning research briefs are a notable part of the service's workflow. They give users a recurring way to monitor selected paper feeds instead of asking an assistant to begin each research session with an entirely new search. The focused-feed model is suited to researchers who want updates organized around particular topics or collections.

Its MCP interface lets an assistant combine discovery, feed-level search, and paper reading in one research flow.