Complete STIQRA manual

Metrics and Feed Stats

Metrics helps you see how reading, filtering, and search habits change over time. Feed Stats helps you see whether the websites you follow are providing new articles reliably. A website's list of article updates is also called a feed (often RSS or Atom).

Both views use information STIQRA already has on your device. Feed Stats can optionally record more details during normal checks, but it does not contact publishers just to create a health score. Status labels use text and symbols as well as colour so you can tell healthy, delayed, blocked, and disabled sources apart.

Metrics dashboard

Metrics opens with a full-width Signal / Noise tile, followed by estimated lifetime, today and this week's savings. Below the highlights, Metrics shows reading activity, filtered impact, the savings breakdown, attention quality, efficiency, signal trends, bandwidth, source health and Top Keywords drill-down figures. Available filtered-keyword statistics follow. These overall figures are directly visible by scrolling. Compact feed summaries open the full per-feed activity measurements.

On iPhone, each Top filtered keywords entry presents Today, This week, and Total as three full-width rows, with the count and a horizontal coloured bar below its single-line label. Each bar compares that period's count with the largest of the three observed counts for the same keyword. A zero count has no coloured fill. These are relative comparisons of observed counts, not progress targets or a historical trend; sorting still follows the selected period.

Feed Stats opens with a source-availability ring and all totals visible in two columns. A small notice appears only while fetch audits are off. The source list uses compact summaries; select a source for its audit-score chart, article totals, availability and technical details, all directly visible by scrolling. TLS policy settings use separate rows.

Fetch-audit controls and the confirmed reset action are under Collection & privacy, below the source list. Recording remains optional and off by default; moving the controls does not enable it. Durable source availability remains visible without fetch audits. Missing measurements stay unmeasured rather than appearing as successful results.

The metrics dashboard summarises productivity-oriented signals such as time saved, Top Keywords drill-down reductions, signal/noise trends, cluster-assisted review, and related reading aggregates.

Use Metrics when you want to understand whether STIQRA is helping you reduce noise, narrow broad result sets, and focus on valuable articles.

Time saved

Time-saved estimates are derived from local assumptions and article activity. Treat them as directional, not exact measurements. They are most useful for comparing trends over time.

This metric exists because STIQRA is designed to help you finish the reading job, not to keep you scrolling longer. Filtering, summaries, source health, saved views, related-story clusters, audio briefings, reports, and Top Keywords drill-downs are meant to reduce wasted scanning time and help you leave with a clearer view.

Where time-saved metrics appear

Time-saved values appear in the Metrics dashboard, related feed-ranking views, and summary surfaces that point you back to Metrics. They are reading-efficiency diagnostics, not social scores, streaks, or pressure to return to the app.

How the estimate works

STIQRA uses a conservative per-item scan-time assumption for local productivity metrics. Current metrics use a default assumption of 20 seconds per skipped or avoided candidate unless an existing local metrics state carries a different value.

The total combines several local signals:

  • Filtered time counts newly hidden article IDs from keyword, language, and local-analysis filters. An article is counted once when it first becomes part of the filtered set, then remembered by ID so repeated refreshes do not keep adding the same item.
  • Focus time counts only the additional articles hidden when you use track-only focus to narrow a broader visible set to tracked matches. It records positive deltas rather than subtracting time repeatedly.
  • Top Keywords drill-down scan time counts qualified cumulative Top Keywords sessions, where at least two Top Keywords taps narrow the same query chain. It estimates avoided scan time from the difference between starting and ending visible candidate counts.

STIQRA clamps negative values to zero and treats these numbers as estimates. They are meant to explain the direction and scale of reduced scanning, not to claim exactly how long you personally would have spent reading.

Privacy of metrics

Metrics are local diagnostics. The time-saved model does not require a remote analytics service. Top Keywords drill-down metrics store aggregate counts such as sessions, tap depth, candidate reduction, and estimated scan time saved; they do not store the raw tapped keyword strings in Metrics.

Reading activity and per-feed usage

Metrics and Feed Stats now include a compact Reading activity view with three local lifetime counters: Items opened, AI reports, and Share actions. Feed Stats shows the per-feed values in each feed row and its detail screen; Metrics also shows these counters in feed rankings. The cards use adaptive columns, so longer titles and values can wrap instead of being cut off on narrow screens.

  • Items opened counts article detail presentations on iPhone/iPad and supported Mac list/card opens, not refreshes, background fetches, marking as read, or time spent reading. It counts repeat opens, not unique articles.
  • AI reports counts distinct report IDs whose local import succeeded after this feature was introduced. Handing off an article for AI analysis does not itself mean a report was generated or imported. A multi-source report counts once in the overall total and once per feed where a report source can be verified against a local article. Imports without a verifiable local source still count in the overall total. Reimporting the same report ID does not increment the count.
  • Share actions counts opening the system article-sharing sheet. The operating system's ShareLink does not reliably tell STIQRA whether sharing was completed or cancelled, so this is not a count of confirmed recipient deliveries.

These activity metrics start accumulating with the updated application, not retroactively. They do not rely on enabling Feed Stats fetch audits, do not make new feed requests, and do not upload reading activity. The totals are lifetime counters, not the current number of reports remaining in the library; deleting a report does not subtract an earlier successful import. A full user-data reset clears the counters.

Top Keywords drill-down metrics

Top Keywords drill-down metrics count cumulative keyword sessions that use at least two Top Keywords taps in the same narrowing chain. STIQRA records aggregate counts such as sessions, tap depth, candidate reduction, and estimated scan time saved. It does not store the raw tapped keywords in Metrics.

A one-tap Top Keywords search can still be useful, but it is not counted as a drill-down session until you tap a follow-on keyword from the narrowed list.

Signal and noise

Signal/noise metrics estimate how much useful content remains after filtering, ranking, archiving, and reading workflows. If the app feels noisy, use these metrics with Filtered Words, Track Keywords, and Source Settings to adjust inputs.

Reading the visual indicators

Metric tiles use the same icons and numbers on iPhone, iPad, and Mac. A filled bar means a measured percentage between 0% and 100%. Counts and durations have no made-up progress target.

Feed Rankings lists only sources in your current refreshable subscriptions that are not disabled in the app's feed preferences. Removing a source removes it from rankings even if historical local counters once existed for it. If there are no eligible sources, rankings say No feed data yet.

In Feed Rankings, Signal is the share of that source's current articles matching your Track keywords. It is not a lifetime score. Signal / Noise compares Track matches with other and filtered articles across your active sources. Each article is counted once, even if several filters match it. Removed sources no longer count, and turning Track-only on or off does not change these numbers.

Without Track keywords there is no Track signal; without enough observations the ratio is unavailable. The small bars compare observed categories, not predictions over time.

Circular audit rings use the optional feed audit score (out of 100); a missing score displays a dash instead of a zero-quality claim. The Feed Stats Healthy ring is different: it shows the percentage of configured sources currently classified as healthy by durable Source availability. The accompanying coloured strip separates healthy, needs-attention, and disabled sources. An empty feed list does not imply 0% health.

Feed Stats dashboard

Feed Stats combines durable source availability with optional per-feed fetch audits from normal refreshes and image loads that already happen. Source availability is the authoritative freshness/availability view; audit scores summarise transport observations and are not the source-health state.

Feed Stats are built from normal refresh behaviour and restart-persistent local observations. STIQRA does not send extra diagnostic feed requests just to explain source health. For each configured source, it can show healthy, stale, temporarily failing, rate limited, feed access restricted (HTTP 401), redirect failure, invalid feed, policy blocked, or disabled; last success and attempt; next automatic check; a separate publisher retry deadline; effective origin; bounded status/category/summary; observed update cadence; automatic versus manual-only mode; and whether last-known-good local articles remain usable.

Feed Stats also helps explain the limits of direct source access. STIQRA contacts enabled RSS/Atom feeds in the current exact selected public-HTTPS initial origins directly, with XMAKYNA out of the normal content path, but many publishers still share CDNs, DNS providers, hosting companies, or bot-defence services. Shared provider hints and repeated source failures can help you notice when several feeds are affected by the same external failure domain.

Feed Stats does not create extra network requests. Durable source health uses redacted refresh observations even when audits are off. Optional audits record data from feed fetches and image work that already occurs during refresh, image prefetch, previews, saves, or visible image loading. Image declaration mismatches are publisher/CDN metadata warnings: STIQRA still trusts validated bytes over a misleading content type or file extension. Low-resolution adaptations counts audited occasions when a feed supplied a valid image too small for the requested large display surface and STIQRA switched to compact presentation rather than visibly enlarging it; a yellow warning explains this behaviour. These adaptations are not counted as image attempts, failures, rejections, or source-availability failures. Image fail counts meaningful image-load failures such as invalid URLs, download/no-data failures, timeouts, and secure image rejection; user-driven cancellations are tracked separately and do not lower the image failure rate. Image/media diagnostics never change source availability.

Feed Stats and the per-source productivity metrics show each source's configured title when available, falling back to a readable label when it is not. Renaming a feed changes the visible name, not the underlying counters, source audit keys or historical results.

Feed Stats controls

Feed Stats includes an Enable fetch audits toggle. When it is off, STIQRA does not record new transport-audit events, but durable Source availability remains available. Turning it on records future fetches that already occur during refresh; it does not refresh feeds by itself and does not create diagnostic-only traffic.

Use Retry now only when Source availability offers it. A rate-limited source waits for the displayed publisher deadline. Other guidance directs you to Source Settings → Subscriptions for URL, redirect, enable/disable, or removal decisions. Use Reset in Feed Stats when you intentionally want to clear on-device fetch-audit history; Reset does not clear durable source health, delete feeds, or remove articles, filters, reports, audio, or preferences.

Why Feed Stats matter

Use Feed Stats when:

  • A source refreshes slowly.
  • A feed frequently fails.
  • You want to compare source reliability.
  • You need to see whether a feed returns no new content or errors.
  • You want to inspect connection and response behaviour at a high level.
  • You want to identify feeds with repeated low-resolution image adaptations, image metadata warnings, image load failures, or secure image rejections.

Interpreting failures

A feed can fail for many reasons: invalid feed data, temporary publisher outage, network problems, TLS issues, redirects, blocked requests, or rate limits.

One failure does not always mean a feed is broken. A single parse/invalid response is temporarily failing; repeated invalid responses can become Invalid feed. A valid 429/503 publisher deadline becomes Rate limited for every known sibling subscription on that origin, including after restart. Manual refresh cannot bypass that deadline, while unrelated origins remain eligible. Existing local articles remain usable unless they were removed by a separate retention or user action.

TLS policy display

Feed Stats can show the active client-side TLS policy. STIQRA uses secure network defaults for external feed and media sessions. Where the operating system reports relevant information, Feed Stats can show protocol and cipher-suite data.

Feed Stats does not claim low-level cryptographic details that the platform does not expose.

Status codes

HTTP status codes can help explain behaviour:

  • 200-level responses usually mean the feed responded successfully.
  • 300-level responses may indicate redirects.
  • 400-level responses often indicate client-side problems such as missing authorisation or invalid URLs.
  • 500-level responses usually mean the server had a problem.

Cache signals

Cache-related headers can explain why a feed returns quickly or does not transfer full content every time. STIQRA uses safe, whitelisted response metadata rather than exposing arbitrary headers.

Metrics privacy

Metrics and Feed Stats are local diagnostics. They are intended to help you understand your own reading and feed performance, not to create a remote analytics profile. Top Keywords drill-down metrics use aggregate counts and do not store raw tapped keyword strings.

STIQRA does not need a remote analytics service to decide which sources are slow, which filters reduce noise, or how much scanning time a local workflow may save.

Time saved, not time captured

STIQRA’s metrics are meant to help you understand and improve your own workflow. They are not engagement analytics for XMAKYNA.

The time-saved view exists because a good reader should give time back. If filters, source health, saved views, and summaries help you avoid scanning the same low-value material repeatedly, the app should show that benefit locally instead of using your reading behaviour to optimise an advertising or engagement loop.

Treat the number as an estimate, not a moral score. The point is not to pressure you to read less or more. The point is to make the utility visible: when STIQRA reduces repetitive scanning, it should help you see that the product is working for you.

Resetting or clearing metrics

Data reset controls in Settings can remove metrics along with other app data when you choose destructive cleanup. Review confirmation text before deleting.

Best practices

  • Use Metrics to tune your workflow.
  • Use Feed Stats to tune your sources.
  • Remove or disable feeds that repeatedly fail or produce low-value content.
  • Use Feed cutoff and filters to keep active views focused.

Clusters and metrics

Clusters does not track session creation, resume counts or reviewed-scope completion, and Metrics no longer has a separate recovery card for that retired feature. Normal explicit article actions still update the relevant reading and Archive state. Fixture-backed clustering quality is developer evaluation, not a personal reading metric.

On iOS, Time Saved uses compact summary rows. Source availability shows source refresh permission separately from global Auto Refresh settings. Eligible now means the source may be checked when scheduling and plan requirements allow; it is not a promised refresh time.