Liminary

About Liminary

Liminary is a research memory layer for professionals whose reputation depends on getting a number right — consultants, analysts, investors and executives who have already done the reading and now need to find it again under time pressure. You save articles, reports, PDFs, AI chat transcripts and videos with one click while browsing, annotate as you read so the capture holds your thinking rather than just the source, and set watchlists per client or project that an agent monitors for new material. The retrieval claim is specific and testable: the vendor states Liminary is 3.4x more accurate than ChatGPT at retrieving the right answer across internal and external sources, publishes its methodology, and traces every insight back to a source that is one click away. It also states plainly that customer data is never used to train models. Beyond recall the product covers the rest of the workflow — fact-checking a claim against what you have saved, spotting gaps in an argument, composing deliverables, and collaborating on a project with up to ten people on Pro or fifty on Team. Magnet, its meeting sidekick, proactively surfaces relevant saved material during a call rather than waiting to be asked. Pro adds MCP connectors and agents plus model selection, which puts the saved corpus in reach of external assistants rather than trapping it in one interface.

Key Features

Tags