A free, local-first Android news reader for Indian and international journalism, with UPSC-oriented study tools.
Today, Explore and Search show only stories with a mapped GS subject and a sufficiently specific syllabus signal. Broad entertainment, celebrity, box-office, reality-TV and routine sports coverage is excluded from these default feeds; stories about an actual law, regulator or institutional policy can still qualify. Saved stories remain available because saving is an explicit user choice.
This is a conservative editorial heuristic over publisher RSS titles, excerpts and section labels—not a validated UPSC relevance score, a PYQ prediction engine or a substitute for reading the primary source. A keyword match alone is not proof of syllabus relevance. Verify important claims against the original public document and the official UPSC syllabus.
The bottom bar has exactly four destinations:
- Today — stories published on the current Asia/Kolkata calendar date.
- Explore — the broader retained news feed, organised by topic.
- Search — a dedicated destination with headline, publisher, topic, subject and GS-paper filters.
- You — the personal dashboard and entry point to Saved, Reading Stats, Quiz & Practice Stats, and Settings.
There is no Archive destination and no Saved bottom tab. Saved is a section inside You. Existing appearance settings remain available from You.
Navigation and Search share one foreground-only 8×8 luminance probe over a full-width 104 dp band at the bottom of the captured page. The region follows the root content height when the IME opens, so adaptive contrast reflects the scene immediately behind the floating controls rather than averaging unrelated content across the whole screen. Sampling retains PrismalAGSL v1.0.4's 600 ms RenderEffect / 1500 ms legacy cadence and adds no dependency.
- The You dashboard shows unique read markers, articles read today, saved articles, submitted quiz attempts, and correct/incorrect answers.
- Legacy read URLs remain read-history markers. They contribute to the unique-article count only; no historical date or reading duration is inferred.
- Reading statistics use persisted foreground reader sessions. Time is counted only while the reader is foregrounded, flushed on lifecycle pause/exit, and grouped by Asia/Kolkata calendar day. Filters cover date range, subject and GS paper.
- Quiz statistics use submitted attempts only. Submitted answer selections, correct answers and provenance are persisted for the latest 500 detailed attempts; aggregate attempt summaries are retained for up to 5,000 attempts. Unsubmitted selections are never stored as attempts.
- RSS metadata classification is an indicative keyword mapping, not a validated UPSC relevance score. It should be checked against the official syllabus and relevant previous-year papers before being used for study decisions.
- RSS supplies attributed headlines, publication dates, images and short excerpts; it does not guarantee full article text.
- CurrentLens attempts best-effort extraction only from publicly accessible publisher pages, without logged-in cookies, subscription tokens, proxying or paywall circumvention. If extraction is blocked, paywalled or incomplete, the original publisher page is offered in the in-app WebView when its HTTPS host matches the publisher allow-list.
- If neither a public article body nor a verified publisher page can be shown, the reader explains that only the RSS headline/excerpt remains available in the feed.
- RSS metadata is cached locally with a 30-day retention window and a 500-article soft limit; saved stories are retained. Extracted public article bodies use a separate cache capped at 30 entries and 24 MB, with a 2 MB per-entry limit and 30-day expiry.
- Newly fetched stories retain the configured originating RSS feed URL in the local metadata cache. Duplicate article URLs keep the first feed match; section-specific feeds are intentionally listed before broad feeds.
- Explicit RSS timezone offsets are preserved. Timezone-less ISO date/time values and date-only values use Asia/Kolkata; unparseable dates remain undated rather than being silently interpreted as UTC.
- Feed URLs and availability can change. The Indian Express publishes an official RSS directory and limits feed consumption to personal, non-commercial use. Publisher terms apply.
- The five-question flow distinguishes headline comprehension, indicative syllabus mapping and original practice prompts.
- Authored practice is not an official UPSC PYQ. A question is labelled as an official PYQ only when its year, paper and official source have been verified.
- The current mapping uses transparent keyword matches over RSS headline/excerpt text. It is preliminary and not an exam prediction.
- No accounts, analytics, telemetry, cloud AI runtime or paid APIs are included.
- Network access is used for RSS feeds and publicly accessible publisher pages. The English dictionary is bundled as a local SQLite database in release builds and works without a runtime dictionary request. Bookmarks, read markers, study sessions, quiz attempts, settings and caches are stored locally.
- The only declared Android permission is INTERNET; cleartext traffic is disabled and Android backup is disabled.
- Runtime dependencies retain their own licences. See THIRD_PARTY_NOTICES.md.
The single concurrency-protected Build Android APK workflow tests the dictionary builder, generates the offline Open English WordNet database from the official CC BY 4.0 dataset, runs JVM unit tests, and publishes a debug-signed APK as a GitHub Release asset. Actions artifacts are retained for 14 days.
Download the latest APK from Releases.
These are testing builds, not production or Play Store releases. Android may ask you to allow installation from your browser or file manager.
Application source: MIT. Third-party libraries and data retain their original licences; see THIRD_PARTY_NOTICES.md.