PapersFlow — an AI research workspace for source-backed work
Discover papers, verify evidence, synthesize findings, and write with citations that stay connected to their sources.
Find and organize research evidence
PapersFlow helps researchers search more than 474 million scholarly works, inspect connected evidence, organize papers into a durable library, and keep citations attached to the claims they support. It is designed for literature reviews, systematic and scoping reviews, evidence synthesis, citation verification, and research projects that need a traceable path from discovery to writing.
Move from sources to source-backed work
Use Doxa for research questions and evidence analysis, Prism for deeper synthesis and counter-evidence, Compose for cited academic writing, and repeatable workflows for structured review tasks. PapersFlow keeps discovery, reading, synthesis, and writing in one workspace so an answer can be checked against the papers behind it instead of becoming detached from its sources.
Connect an AI agent through MCP
Agents can call the hosted PapersFlow MCP server over Streamable HTTP. OAuth and personal access tokens use narrow scopes for public research, private-library reads, library writes, research-result reads, and paid research runs. Start on the Free plan without a payment card, then request only the permissions the agent needs. The developer resources describe the endpoint, authentication flow, scopes, and machine-readable manifests.