Web UI

← Home

lit-monitor first-run   # interactive first-time setup, then launches the server
# or, once credentials are configured:
lit-monitor serve

first-run prompts for write-once credentials and your server host/port, then starts the UI. On subsequent launches use lit-monitor serve.

The server binds to 127.0.0.1 by default. Pass --host 0.0.0.0 for LAN access; the server has no authentication of its own. A persistent health badge in the nav polls /api/health and shows pipeline status at a glance. An optional developer page is mounted at /dev when you start the server with lit-monitor serve --dev.

Set a schedule — that's what makes it a monitor

The single highest-value thing to do after setup is install a recurring schedule from the /schedule page (launchd / systemd). With it, new papers arrive already searched, ranked, extracted, and filed between visits — see Scheduling — the monitor loop.

Setup wizard

Open http://127.0.0.1:8765/setup in any browser. An 8-step wizard covers credentials, paths, extraction config, topics, domain context, theme vocabulary, tracked researchers, and item routing, with live credential checks at each step.

To configure the same settings by hand instead, see Configuration.

Dashboards

After setup, the dashboards take over. The top nav groups the pages by what you're doing — Run, Ask, Explore, Tune, Setup.

URL	Group	What
/discovery	Run	Latest discovery run summary, run history, per-paper cards with one-click relink / re-extract actions. Run-now / Dry-run / Stop buttons.
/brain-build	Run	Extract your existing Zotero library into the index. Progress bar, per-paper table, recent runs. Start / Stop / Resume. Live JSONL log stream.
/schedule	Run	Install or remove a recurring schedule (launchd / systemd).
/ask	Ask	Ask natural-language questions of your corpus (the lit-monitor ask pipeline in the browser): prose answer + results table, with show/edit Cypher, recent-questions history, and save-to-vault. Requires the knowledge graph (lit-monitor graph backfill --all).
/corpus	Explore	A read lens on every processed paper in the state DB: searchable list with a theme filter, plus a per-paper detail view (extraction, score, knowledge-graph, related work, Zotero / Obsidian links, relink / re-extract). Related work works without a knowledge graph (vector similarity, each result labelled by source) and the Obsidian link is a clickable deep-link into your vault. Not a Zotero replacement — it reads what the pipeline already processed.
/graph	Explore	A read-only, corpus-wide overview of the knowledge graph: entity counts by type, edges by predicate, entity counts by NER source, and graph-indexed coverage. Read lens only — running a graph backfill is a CLI action.
/themes	Explore	Review the theme clusters your library has been grouped into: per-theme paper counts and a drill-in to the papers in each theme.
/trending	Explore	The trending-concept suggestion queue — concepts rising in your recent library activity, with accept / dismiss actions.
/domain	Tune	Edit the free-text domain-focus paragraph and run the LLM extraction that turns it into the structured domain_context ranking signal.
/insights	Tune	A read-only window into the active-learning engine: the current interest-vector state with top-aligned papers, per-cluster atrophy weights, the feedback mix by signal and source plus its timeline, and recent feedback events. Read lens only — feedback is still captured through the discovery and themes buttons.
/settings	Tune	Advanced settings — edit ranking weights, clustering, embeddings, and discovery/delivery config sections from the browser.
/setup	Setup	The 8-step onboarding wizard (see above).