Arivu's New Direction: A Self-Hosted Second Brain

Arivu began as a way to stop useful links from disappearing into a bookmark archive. That problem still matters, but saving and summarizing links is only one part of building knowledge.

The current Arivu is a self-hosted second brain and private knowledge workspace. It brings bookmarks, notes, annotations, imported material, links, search, a knowledge graph, and evidence-backed Insights into one system that you operate.

Its core loop is simple:

1. Capture a link, note, quote, or file without designing a taxonomy first.
2. Connect ideas with explicit links and backlinks, then inspect locally derived concepts, entities, and sources.
3. Discover material with full-text search, cited Ask responses, or a focused graph.
4. Learn by revisiting useful material and examining patterns with their supporting evidence.

Five places, one knowledge system

Arivu now has five primary destinations.

Home is the daily knowledge pulse. It brings together the daily note, active work, new material, useful memories, and supporting Focus, Review, and Board views.

Library is the unified collection. It leads with bookmarks, notes, daily notes, annotations, and knowledge objects. Generated concepts and entities remain available in a separate derived view and in Graph instead of competing with material you saved or wrote.

Notes is a first-class writing workspace. A note can stand alone, link to saved material, receive backlinks, and carry supporting tasks or reminders.

Graph shows typed relationships in a bounded, focusable view. Provenance and confidence make derived connections inspectable, and a list alternative keeps the information usable without relying on a visual map.

Insights shows deterministic learning patterns such as emerging themes, recurring connections, forgotten value, and knowledge gaps. Each pattern points back to the material that produced it.

Capture and Search / Ask remain available across the product. Supporting workflows such as Inbox, Focus, Review, Board, imports, exports, the browser extension, and the CLI are still present. They help the main loop without defining the navigation.

AI is optional

Arivu does not require a model provider to be useful. Capture, notes, explicit links, backlinks, local extraction, full-text search, graph structure, and deterministic Insights work without one.

An optional provider can improve summaries, embeddings, explanations, and synthesis. Embeddings add similarity relationships when configured; they are not required for the graph or for ordinary search. Ask builds cited responses from material already saved in Arivu and keeps the sources visible.

That distinction matters. The knowledge system should not stop working because an external model is unavailable.

Ownership is part of the direction

Arivu’s core runs as a single Go binary backed by SQLite. On supported hosts, the installer can separately manage an isolated native runtime for complete background capture. Data can be backed up and exported, and existing browser extension, CLI, import, export, and administration workflows remain intact.

Self-hosting is a tradeoff: you operate the instance and its backups. In return, the workspace, database, and provider choices stay under your control.

This is a sharper product than the earlier all-purpose workflow. Arivu is not trying to become a social network, a team collaboration suite, or a full project manager. It is focused on helping an individual turn captured material into connected, retrievable knowledge.

See the current feature map or start with the Capture and Library guide.