Product tour

Turn papers into a structure you can inspect

See how Agent Bayes helps you map a literature, trace its terminology, develop an argument over time, and check substantive claims against the sources behind them.

Literature exploration

Map a literature without smoothing out the debate

Bring a focused corpus from Zotero or upload papers to a knowledge base. Then shape an editable mindmap over multiple turns: give a focused instruction, inspect the citation-backed claims Agent Bayes adds, and use follow-ups to narrow, expand, or correct the map.

Related findings cluster into branches, while competing positions remain separate and traceable to their sources. You get a visible structure you can question, reorganize, and use to decide where deeper reading matters.

  • See themes and disagreements across a bounded corpus
  • Keep each substantive claim connected to its sources
A demo literature map built from a bounded archaeology corpus, with source citations kept on the claims they support.

Argument development

Develop an argument as the evidence changes

Research rarely keeps the outline it started with. Use the mindmap as a working structure for a chapter, review, or proposal. Ask Agent Bayes to deepen a thin branch, compare methods, or bring new papers into the existing map.

You can rewrite nodes and reorganize branches while the version history preserves what changed. The argument remains yours, and the structure does not disappear when the conversation ends.

  • Expand one branch without replacing the surrounding structure
  • Return to an editable map instead of reconstructing a chat
A focused research request adds cautious new branches while the surrounding argument stays visible and editable.

Citation inspection

Open the citation and read the source

Each substantive mindmap claim keeps its citations close. Select a node, open its indexed citations, and read the passage behind the claim without leaving the map.

Check the wording in context and decide whether the source supports the claim, calls for a narrower interpretation, or belongs elsewhere in the argument.

  • Read the passage attached to a selected claim
  • Keep your judgment between the source and the draft
The selected claim and its attached source text remain visible together for inspection.

AI verification

Test a node against its attached citations

Select a mindmap node and run AI Verification. Agent Bayes reads the passage text from every attached citation, scores how closely the node's claims align with those passages, and lists specific discrepancies.

The result appears as a new subtree beneath the original node, with four alternative phrasings to consider. The original stays unchanged.

  • See which phrases the attached passages do not support
  • Review alternatives without changing the original node
An AI Verification result showing citation backing, a specific discrepancy, and four alternative phrasings.

Terminology exploration

Trace the vocabulary your corpus actually uses

Fields rarely use one stable vocabulary. Build a terminology graph for a knowledge base to see its domain terms and which concepts appear together in the same semantic passages.

Load a term and follow its neighbourhood. Select a node to read passages that mention it across the corpus, ordered by how informative they are about the term. Select an edge to see passages where both connected terms appear together.

  • Follow related concepts through a bounded corpus
  • Open the passages behind each term or connection
A local neighbourhood around Palace 6000. Node size reflects paper count, while connection width reflects co-occurrence across papers.

Your corpus, your judgment

See what Agent Bayes looks like with your literature

Agent Bayes works from the sources you choose. It can organize, compare, and help verify evidence. You decide what belongs in the argument.

Join the Waiting List