Structure before inference
Treefiny reads reliable structure—citations, headings, links, authors, dates, and user-authored relationships—before asking AI to interpret meaning. Models propose what cannot be extracted directly.
03 / Intelligence
Do not just retrieve an answer. Reveal the path.
Treefiny combines reliable structural extraction with bounded semantic inference—keeping source facts, user authorship, rules, and AI proposals visibly distinct.
Join the private alphaInspectable graph intelligence
Ask, trace, compare, and explain.
The graph can answer questions, trace relationships, reveal useful neighborhoods, and show what changed without hiding the objects and paths beneath fluent prose.
Every edge tells the truth
Connections distinguish what you authored, what a source states, what a rule derived, and what AI inferred. Certainty and review state stay separate, visible, and correctable.
Ask, trace, explain
Ask a question against a focused subgraph, trace how two ideas connect, or explain why a thought matters. Answers return inspectable objects and paths—not only fluent prose.
Find neighborhoods and bridges
Discover thematic regions, central ideas, surprising cross-field bridges, orphaned thoughts, and ambiguous links. Graph signals guide attention without pretending connectivity equals truth.
See what changed
New and updated sources produce a reviewable graph diff. Treefiny shows what was added, contradicted, relinked, or superseded without silently rewriting permanent thoughts.
Retrieve with a visible path
Lexical search, semantic similarity, metadata, time, and graph traversal work together. Every item in an AI context packet explains why it was selected and how it reaches the active question.
This architecture learns from the open-source Graphify project, particularly its structural-first graph construction, explicit relationship provenance, community analysis, incremental updates, and query, path, and explain operations. Treefiny adapts those principles to human learning and thinking rather than codebase mapping.
AI with bounded authority
AI is the interface. The graph is the memory. You are the author.
Treefiny can ask, retrieve, compare, challenge, and draft. It must never manufacture a source, silently rewrite the graph, or hide uncertainty behind a confident answer. Useful intelligence begins with legible boundaries.
Every change is reviewable.
Accept, edit, dismiss, defer, disable, or undo AI-proposed connections and content.
Every answer shows its footing.
User statements, quoted sources, system inference, and speculation remain visibly distinct.
Knowledge stays within its boundary.
Private graphs are scoped, sharing is opt-in, and export, deletion, and model-use policies are explicit.
Reflection wins over engagement.
The product optimizes for clarity and durable understanding—not streaks, noise, or endless capture.
The adjacent landscape
Learn from strong products. Keep the thesis distinct.
Treefiny sits beside connected-note systems, visual research spaces, AI notebooks, and learning tools. Each proves an important behavior; Treefiny joins those lessons in one longitudinal reasoning graph built around provenance, agency, recall, and creation.
Capacities
Connected, typed objects and flexible views.
Treefiny extends the object model into explicit reasoning relationships, source boundaries, recall, and claim-level synthesis lineage.
Explore ↗Heptabase
Visual learning and research through cards, sources, and whiteboards.
Treefiny keeps every visual arrangement synchronized with a typed graph, programmable reflection, memory, and reviewable AI patches.
Explore ↗Tana
Structured nodes, supertags, AI, commands, and events.
Treefiny puts evidence, uncertainty, relationship provenance, calm reflection, and reversible authorship at the center of automation.
Explore ↗NotebookLM
Source-grounded AI research, citations, and multimodal explanations.
Treefiny develops durable thought objects and cross-project paths that connect sources to a person’s evolving ideas, decisions, and creations.
Explore ↗RemNote
Notes, source annotation, AI study tools, flashcards, and spaced repetition.
Treefiny connects recall to questions, claims, evidence, projects, conceptual transfer, and the wider reasoning graph.
Explore ↗Obsidian
Private local Markdown, links, graph and canvas views, and extensibility.
Treefiny pairs durable ownership with an opinionated AI-first thinking method, typed provenance, and explainable graph operations.
Explore ↗Comparisons reflect each product’s public materials as reviewed on August 26, 2026. They describe emphasis rather than exhaustive capability or private roadmaps.
The build standard
Fast enough for a fleeting thought. Rigorous enough for a serious conclusion.
A great Treefiny experience makes capture effortless, relationships understandable, provenance inspectable, recall purposeful, and AI actions reversible. It works with keyboard, touch, and assistive technology; remains calm from phone to wide screen; and turns the same knowledge graph into the right view for the moment.
Private alpha
Bring your thinking with you.
Join the Treefiny waitlist for product previews, research invitations, and early-access openings. This list is only for Treefiny.