Track blood glucose with meals and exercise

PlateLens keeps each blood glucose reading at the time it was observed. You can add a reading yourself or let the phone sync readings you have allowed Apple Health or Health Connect to share. Timed meals and exercise logged in PlateLens appear as context on the same day without claiming that one caused the other.

Log a reading manually

1. Open +, choose Measurements, then Blood glucose.
2. Enter the value in the unit shown. On mobile, tap the unit beside the value to switch between mg/dL and mmol/L for this entry.
3. Set the date and the time when the reading was taken. PlateLens uses the selected Diary day when you opened + from a past day.
4. Optionally add meal context, sample type or a private note.
5. Save the reading.

Your glucose display unit is a separate preference from metric or imperial weight and distance. PlateLens initially chooses the unit commonly used in your account country. On mobile, change the account preference in More → Preferences → Units → Blood glucose. That preference controls how readings appear in mobile history, the web app, Watch and PDF reports.

The unit beside the value in the mobile reading form applies only to that entry. Switching it converts the unfinished value without changing your account preference. PlateLens stores one canonical mmol/L value; choosing a different input or display unit does not create a second reading or change an existing reading.

Sync from Apple Health or Health Connect

On the phone, open Health Integration and allow Blood Glucose read and write access. PlateLens imports the readings the health platform makes available. Manual PlateLens readings created on the phone, web app or Watch are mirrored by a consenting phone the next time it is active. A continuous sensor may contribute frequent samples; a finger-stick meter may contribute only one or two readings.

• The first glucose sync reads up to the previous 30 days. Later syncs request changes when PlateLens opens or returns to the foreground.
• The time shown is the reading's observation time, not the time your sensor app eventually delivered it.
• PlateLens prevents the same health record from being imported twice and applies explicit updates and deletion events supplied by the health platform. It never treats an empty or partial history read as proof that records were deleted.
• Permission changes belong in Apple Health or Health Connect. Turning permission off stops future access; it is not a delete action and readings already copied to your PlateLens account remain.

Where readings appear

Use the History charts

Day is a 24-hour timeline with clock labels, glucose values on the vertical axis and a legend for the kinds of readings and context that are actually present. Tap, click or drag across the chart to select the nearest observation and see its exact value, time, source and measurement method. Select a meal or exercise marker to see the event at that time; an exercise with a recorded duration also shows its observed interval. Keyboard and screen-reader actions can move through the same observations, and the exact ledger remains available below the chart.

7 days, 30 days and 90 days change the horizontal axis to dates. Each observed day shows its median reading and its lowest-to-highest observed span. A day with one reading collapses to one point, and a day with no readings stays a visible gap rather than borrowing a neighboring value. Select a day to open its 24-hour timeline and meal/exercise context. All keeps the complete chronological ledger without compressing an unbounded history into a misleading chart.

Points, sensor paths and real gaps

One reading is always one point. Two intermittent meter readings remain two points. PlateLens draws a line only across a sufficiently dense sequence from the same contributing app/source stream and stored measurement method. A gap longer than 20 minutes starts a new segment. It never joins records whose stored method identifies one as capillary and another as interstitial, invents samples between observations, or stretches the first or last sensor value across the rest of the day. If a source supplies no method, PlateLens keeps it explicitly Unknown rather than guessing.

Manual readings are visually distinct from health-synced readings. PlateLens does not calculate a clinical average, time in range or a meal-response score from sparse data because those summaries would imply coverage the account may not have. The Analysis view calls sparse data Limited, keeps its individual facts visible, and waits for repeated observations before describing a pattern.

Analyze what was observed

On mobile, open Glucose from Home, Diary or More and choose Analysis. Use its compact controls to choose the period and measurement method, and expand the information section when you want more explanation. On the web, open Progress → Blood Glucose → Analysis. Choose 7, 30 or 90 days. Both surfaces show the same server-derived answer in the account's display unit; changing between mg/dL and mmol/L changes presentation only.

• Observed overview shows the reading count, days with readings, median, and lowest-to-highest observed value. These are descriptions of stored readings, not a target range.
• Measurement method keeps interstitial sensors, capillary finger-pricks, other sample types and unknown-method readings in separate stored groups whenever the health source supplies that evidence. If a period contains more than one group, All methods remains an inventory and asks you to choose one before showing comparisons.
• How this range was observed separates dense sensor segments from independent spot readings, names the source mix, and counts only observed sensor coverage. Gaps longer than 20 minutes and overlapping sensors never become extra observed time.
• By time of day groups readings into overnight, morning, afternoon and evening using trustworthy local-time evidence. Each displayed median gives every observed day equal weight and appears only after at least three readings across three days. A reading without local-time evidence stays unclassified rather than being guessed into a band.
• By logged meal context groups only the context stored on the glucose reading, such as fasting, before a meal or after a meal. It uses the same per-day weighting and never assigns an untagged reading automatically.
• Around logged meals and exercise compares observed glucose before and after timed meals and timed exercise logged in PlateLens with readings in both required windows. It shows an aggregate only after at least five unique reading pairs across three days, and reports the observed minimum-to-maximum change as well as the day-weighted median.

If a 90-day window contains too many dense sensor observations to analyze in one request, PlateLens offers a 30- or 7-day window. It never returns a hidden sample or partial analysis. History and account export retain every reading. PDF reports summarize the full observed coverage, preserve every spot reading, and may sample only dense CGM points in the printed table.

Ask your AI Coach about glucose

Open Coach on mobile, from the web sidebar, or in the Apple Watch app and ask about glucose already stored in your PlateLens account. There is no separate Coach switch to activate: glucose becomes available as context after the account has a reading, while Apple Health or Health Connect permission remains a separate phone setting.

• For one date, ask what was observed and how readings line up in time with confirmed meals or PlateLens-logged exercise. The Coach reads that exact day; it does not substitute the latest reading from another date.
• For repeated observations, ask about the previous 7, 30 or 90 days. The Coach uses the same bounded analysis as the Analysis screen, including sufficiency, observed coverage, source mix, reported measurement method and eligible meal or exercise timing context.
• The Coach distinguishes interstitial sensor observations from capillary or other spot readings only when the stored source reports that method. Unknown stays unknown, and sparse readings never become an invented continuous curve.

The Coach can read this context only when it is relevant to your conversation. It cannot create, edit or delete a glucose reading, request health permission or start a sync. Its glucose context excludes private glucose notes, record identifiers, native provider identifiers and source-app names.

Share readings with a connected AI assistant

The optional PlateLens MCP connection lets a compatible assistant read glucose already stored in your account. This has its own permission, glucose.read, separate from your phone's health permission and the PlateLens Coach.

1. Add https://mcp.platelens.app/mcp as a custom MCP connection in your assistant.
2. Sign in at auth.platelens.app with the Google or Apple account you use in PlateLens.
3. Review the requested permissions and explicitly select glucose. Existing connections must reconnect to grant this additional permission.
4. Ask for your stored readings on an exact date or for a range of up to 31 days. For example: “Show my recorded glucose readings for September 1, including their units, times and measurement methods.”

The assistant receives at most 100 observations per page, with the stored mmol/L value, observation time and recorded timezone context, source platform and reported measurement method. It can request further pages. Private notes, source-app names and native/provider identifiers are excluded. A date with no stored observations returns no readings; the connection does not start a phone sync or replace missing values with zero.

The connection cannot add, edit or delete readings, set clinical targets or provide glucose monitoring or alarms. To stop sharing, sign in at auth.platelens.app and revoke the connection. Use this browser manager as well if a client cannot load its connection list. Revoking access stops future reads; data already returned to an assistant is governed by that provider's settings and policies.

Meal and exercise context

Meal time matters because a meal logged later can still sit at the time it was eaten. Exercise keeps its confirmed start and duration. Timelines mark the confirmed start; on the web, Diary and Blood Glucose History also show the timed event below the curve with the meal's calories and main macros or the exercise's full interval, duration and calories. These events sit beside observed glucose only as temporal context: food, supplements, sleep, stress, medication and many other factors can affect a reading.

With alcohol tracking on, a meal event that includes drinks with an estimate also shows its approximate alcohol, such as ≈19.7 g of alcohol, on mobile, on the web and in PDF reports. It is context for the same timeline, not a prediction of how a reading will change.

Those facts are projected from the original meal or exercise. They are not copied into glucose records, so correcting the source item updates the same timeline without creating conflicting data.

PDF reports and account export

Weekly and monthly PDF reports automatically include Blood Glucose when the report period has readings. Across multiple days, the chart uses each day's observed median and lowest-to-highest span and leaves dates without readings blank. Every spot reading remains in the detailed table. Dense sensor coverage is shown with real gaps, accompanied by the complete reading count and observed coverage rather than a clinical interpretation; only dense CGM points may be sampled in the printed table for readability. The glucose section can appear even on a day without a meal.

Your account-data export includes every canonical glucose reading in the same private JSON file as the rest of your account. Large exports are prepared in the background: you can leave the screen and return to Export My Data to resume the same preparation. Changes made while the file is being generated can affect its contents. Internal deduplication records, provider-native identifiers and synchronization bookkeeping are not included.

Time zones and daylight saving time

A reading always stores a real instant. When a manual entry or health source supplies trustworthy occurrence-time zone or offset evidence, PlateLens also preserves the civil day and clock time where it happened. If a provider omits that evidence, Analysis leaves the reading out of local-time bands instead of guessing. A manual clock time skipped by daylight saving time cannot be saved; when an hour occurs twice, PlateLens lets you choose the intended occurrence.

If something is missing

• Confirm that the reading exists in Apple Health or Health Connect and that PlateLens has Blood Glucose read access.
• Open PlateLens on the phone and leave it in the foreground long enough to complete a sync.
• Remember that some sensor providers send health-platform readings later than their observation time.
• If one source has a long gap, PlateLens leaves that gap instead of joining it to a different source.
• If Analysis asks for a shorter window, choose 30 or 7 days. Your longer History is still intact and no readings were discarded.
• To edit or delete a manual reading on mobile, open Blood Glucose History, tap the reading, then open …. A left swipe is only a delete shortcut. On the web, use Progress → Blood Glucose → History.

Read the product story in Blood glucose, meals and exercise in one timeline. Still need help? Contact PlateLens Support.