● Food Intelligence · Vegetable

Celery Glycemic Index and Calculate Your Own Glucose Response

Across 19,889 celery-containing meals, members averaged +26.3 mg/dL (median +22). With a published GI of 15 and GL of 0, celery is ~1g of carbs per stalk — almost entirely water and fiber. The cohort signal traces to what celery rides with: in soup or mirepoix, +31.7 mg/dL (+5 mg/dL vs. celery-alone); with ranch dip, +28.4 mg/dL (+2 mg/dL); with peanut or almond butter, +24.1 mg/dL — actually below the celery-alone baseline.

GS
Reviewed by Grace Shryack
Signos Proprietary Data·Updated May 20, 2026·10 min read

● Powered by Signos cohort dataHigh confidence · n=19,889
The swap calculator below draws on 19,889 matched-pair logs — a broader cohort than the page's single-item primary, used to give every ingredient swap statistical power. Welch's t-test on matched pairs, not third-party glycemic-index tables. Each swap shows its sample size and confidence tier inline.
● FREE SIGNOS TOOL

Build Your Own Low GI Recipe

Powered by 19,889
Signos food logs

Tap any ingredient below to swap it for a glucose-friendlier alternative. Your Signos Glucose Score, predicted curve, and nutrition update in real-time. Every swap is data-backed by real CGM response measurements.

Ingredients — Tap to Swap
Meal Context
Pre-meal sequence
Activity after meal
Time of day
36
of 100
Signos Glucose Score
Moderate Spike Risk
75% of members fall between +22 and +32 mg/dL
Predicted Curve · 3-hr window
1401001201401601800m30m60m90m120m150m180m+27
Peak
+27
mg/dL
Time in Range
100%
3hr window
Above 140
0m
danger zone
Swaps
0
applied
Want your real response?
Predictions are cohort averages. Wear a CGM to see your actual response.
Get a Signos CGM →
● Key Findings · Does celery spike blood sugar?

Almost never — celery has a GI of 15 and GL of 0, and the +26.3 mg/dL cohort mean across 19,889 logs reflects what celery rides with, not the stalks themselves. Across 19,889 celery-containing meals from 5,021 unique members, the median glucose peak was just +22 mg/dL and 45.0% of all responses landed in the low-spike range. Single-item celery logs averaged just +24.6 mg/dL across 754 entries. The pairings the data actually surfaces are mild: with peanut or almond butter, the peak runs +24.1 mg/dL (−2 mg/dL below the celery-alone baseline of +26.4 mg/dL across 3,245 logs); with ranch or blue-cheese dip, +28.4 mg/dL (+2 mg/dL); in tuna or chicken salad, +25.9 mg/dL (essentially flat); in soup or mirepoix, +31.7 mg/dL (+5 mg/dL — the broth, bread, and roux, not the celery).

  • ▸The carb-load effect dominates: meals with ≥60g of carbs alongside celery ran 58.4% higher than the low-carb baseline (n=1,878 vs n=14,967, p<0.001).
  • ▸Celery + peanut butter is net-negative: the pairing averaged +24.1 mg/dL — 2 mg/dL below the celery-alone baseline of +26.4 mg/dL across 3,245 logs (high confidence).
  • ▸Soup and mirepoix pairings drive the biggest delta: celery in soup averaged +31.7 mg/dL across 1,293 logs — a +5 mg/dL jump over the celery-alone baseline, almost entirely from the broth, bread, and roux.
  • ▸Snack-time celery is the cleanest occasion: snack logs averaged +23.2 mg/dL across 3,222 entries — the lowest meal-time slot for any celery context.
● Signos Rules

Three citable insights from this recipe's data

Rule 1
"The Peanut-Butter Pass"
−2 mg/dL vs. alone
Celery + peanut or almond butter averaged +24.1 mg/dL across 3,245 logs — 2 mg/dL below the celery-alone baseline of +26.4 mg/dL (high confidence). The classic snack — "ants on a log" with the raisins removed — behaves cleanly in the cohort. Peanut butter contributes fat and protein that slow gastric emptying, and the small snack-size portion doesn't pile on enough carbs to spike. It's the rare pairing in the atlas where adding something to a food brings the peak down.
Rule 2
"The Mirepoix Penalty"
+5 mg/dL vs. alone
Celery in a soup or mirepoix base averaged +31.7 mg/dL across 1,293 logs — a +5 mg/dL jump over the celery-alone baseline of +26.4 mg/dL (high confidence). The driver isn't the celery; it's the soup format itself: a flour roux to thicken the broth, bread or crackers eaten alongside, and the larger overall meal context. The form-anchor split agrees — diced-cooked celery averaged +30.3 mg/dL versus +26.2 mg/dL for raw stalks (n=367 vs 12,350).
Rule 3
"The Snack-Time Edge"
+23.2 mg/dL
Snack-occasion celery logs averaged +23.2 mg/dL across 3,222 entries — lower than breakfast (+25.1), lunch (+26.5), or dinner (+28.8). The format is the lever: snack-time celery is typically eaten alone, with peanut butter, or with a small dip — none of which add meaningful carbs. Lunch and dinner celery rides with sandwiches, soup-and-bread, or salad-with-crackers — and the carb companion is what moves the number up.
● Curious about your own?

Curious how your body responds to celery specifically?

Across 19,889 logged meals containing celery, the average glucose peak was +26.3 mg/dL — but the IQR spans 23 mg/dL (p25: +11, p75: +34), meaning individual responses vary enormously. Your own result depends on meal context, dip choice, and metabolic state. A CGM tells you which side of that range you actually land on.

Learn how Signos works
Why this meal spikes

This meal stacks 3 independent spike drivers — together they account for +26 mg/dL.

Driver 1
Meal-level carbohydrates (≥60g when paired)
+13.9 mg/dL
The dominant lever: meals with ≥60g carbs ran 58.4% higher than low-carb logs (n=1,878 vs n=14,967, p<0.001). Celery itself is GI=15 with GL=0 — every point of that 58.4% comes from the bread, crackers, or soup-thickener around it.
Driver 2
High-sugar pairings (≥20g sugar)
+9.1 mg/dL
Meals containing ≥20g sugar alongside celery spiked 37.9% higher than the low-sugar baseline (n=2,445, CI: 8.27 to 9.91 mg/dL, p<0.001). Sweetened dips, fruit-and-celery snack plates, and the occasional sweetened broth drive most of this effect.
Driver 3
Individual glucose variability (IQR 23 mg/dL)
+11.5 mg/dL
The interquartile range spans 23 mg/dL (p25: +11, p75: +34 mg/dL), meaning personal metabolic state — insulin sensitivity, microbiome, sleep, stress — accounts for a substantial portion of any single reading.
● Which bucket are you in?

Some members spike +23.8 mg/dL. Others spike +42.7. The only way to know how you'll spike is to measure with a Signos CGM.

See Signos Plans
S
● What surprised us
What surprised us about celery is the peanut-butter pairing. The 1990s diet-lore framing was that celery cancels calories — physiology says no, but the cohort says something more interesting: celery + peanut butter actually runs below celery alone (+24.1 vs +26.4 mg/dL across 3,245 matched logs). It's the rare pairing in the atlas where adding something to a food brings the peak down — the fat and protein in the PB slow gastric emptying, and the snack-size portion doesn't pile on enough carbs to outpace that. Celery is the cleanest example we have of a food whose published GI and cohort behavior almost perfectly agree.
— Signos Data Science Team

Why this happens, physiologically

Celery has a published glycemic index of 15 and a glycemic load of 0 — a 2-stalk serving contains roughly 2g of carbs, almost all of it fiber, with the bulk of the mass being water (~95% by weight). The +26.3 mg/dL cohort mean across 19,889 logs is therefore not a measurement of the celery; it's a measurement of the snacks, dips, and dishes celery tends to live alongside. The data confirms this directly: meals with ≥60g of carbs alongside celery ran 58.4% higher than the low-carb baseline (n=1,878, p<0.001), and the carb-bucket analysis shows a near-linear escalation from +23.8 mg/dL in light (0–40g) meals all the way to +42.7 mg/dL in very-heavy (110g+) meals — a 79.4% jump. The peanut-butter pairing pulls in the opposite direction: it averaged +24.1 mg/dL across 3,245 logs, 2 mg/dL below the celery-alone baseline, with the fat and protein in the PB slowing gastric emptying. The mechanism is observational, not yet causal, but consistent with the broader pattern of fat-and-protein pairings damping post-meal glucose excursions.

● Three mechanisms explain celery's glucose range

Meal carb load, fat-and-protein anchors, and dish format each drive measurable shifts in celery's glucose response

  1. Mechanism 1
    +58.4%
    High carb co-load
    Meals logged with ≥60g total carbs showed peaks 58.4% above the low-carb baseline (n=1,878, p<0.001). Celery contributes ~1g of carbs per stalk — every point of that 58.4% comes from the bread, crackers, or soup-thickener around it.
  2. Mechanism 2
    −2 mg/dL
    Peanut-butter pairing
    Celery + peanut/almond butter averaged +24.1 mg/dL vs. +26.4 mg/dL alone (n=3,245, high confidence). The fat-and-protein anchor of the nut butter, plus the snack-size portion, brings the peak below baseline — the only common celery pairing that does.
  3. Mechanism 3
    +5 mg/dL
    Soup / mirepoix format
    Celery in a soup or mirepoix base averaged +31.7 mg/dL vs +26.4 mg/dL alone (n=1,293, p<0.001). The broth, the bread on the side, and the flour roux drive the +5 mg/dL delta — the celery itself is along for the ride.
● Fit Check
Celery is one of the lowest-glycemic vegetables in the atlas — and the rare food where adding peanut butter actually lowers the cohort peak.
✓ This is for you if
  • ▸You snack on celery and peanut butter. That pairing averaged +24.1 mg/dL — 2 mg/dL below the celery-alone baseline across 3,245 logs, the cleanest pairing in the cohort.
  • ▸You eat celery at snack-time. Snack-occasion celery logs averaged +23.2 mg/dL — the lowest meal-time slot, across 3,222 logs.
  • ▸You keep total meal carbs under 40g. Light-carb celery meals averaged just +23.8 mg/dL — the lowest carb bucket in the cohort (n=14,967).
  • ▸You dip raw celery in ranch or blue cheese. That pairing averaged +28.4 mg/dL — just +2 mg/dL over the celery-alone baseline across 1,690 logs.
✗ Not for you if
  • ▸You eat celery primarily in soup or mirepoix-based dishes. That pairing added +5 mg/dL over the celery-alone baseline across 1,293 logs (treatment mean +31.7 mg/dL).
  • ▸You build celery into a 110g+ carb meal. Very-heavy-carb meals with celery averaged +42.7 mg/dL — 79.4% above the light-carb baseline.
  • ▸You rely on the diced-cooked form (mirepoix, stew base). That form averaged +30.3 mg/dL across 367 logs — a +4.2 mg/dL delta over raw stalks.
  • ▸You drink celery juice as a wellness ritual expecting zero impact. Celery juice averaged +27.6 mg/dL across 663 logs — slightly above raw stalks, likely from the concentrated fructose-and-glucose hit of juicing.
● How it fits your day

Calorie band and pairings (member-measured)

Per serving
10–1,270 kcal
Ranges from a 2-stalk raw celery snack (~15 kcal) to a soup-and-bread celery meal (~1,270 kcal in the very-heavy carb bucket); single-item celery logs averaged 110 kcal — among the lowest in the atlas.
● Pair before
  • ▸Pair celery with peanut or almond butter when you need a snack: that pairing averaged +24.1 mg/dL — 2 mg/dL below the celery-alone baseline across 3,245 logs.
  • ▸Keep total meal carbs under 40g: light-carb celery meals averaged +23.8 mg/dL vs. +42.7 mg/dL in the 110g+ carb bucket — a 79.4% escalation.
● Pair after
  • ▸Time a 15–20 min walk after celery-in-soup meals, which averaged +31.7 mg/dL — the highest pairing in the cohort and usually accompanied by bread or crackers.
  • ▸If celery rides in a multi-course dinner, note that snack-time celery logs averaged +23.2 mg/dL vs. dinner at +28.8 mg/dL — favor between-meal timing when possible.
● Avoid pairing
  • ▸Avoid relying on celery in a flour-thickened cream soup as the low-spike choice — the soup-mirepoix pairing added +5 mg/dL over baseline (treatment mean +31.7 mg/dL, n=1,293).
  • ▸Avoid stacking celery into a 110g+ carb meal expecting the celery to lower the spike — very-heavy-carb celery meals averaged +42.7 mg/dL, well above the celery-alone baseline.
● Quick definitions (click to expand)
mg/dL — milligrams per deciliter. The unit blood glucose is measured in. A rise of "+30 mg/dL above baseline" means blood sugar went up 30 units after the meal.
Glycemic Index (GI) — a 0–100 score for how fast a food raises blood sugar in lab tests. Under 55 = low, 56–69 = medium, 70+ = high.
Glycemic Load (GL) — GI adjusted for portion size. Under 10 = low, 10–19 = medium, 20+ = high.
CGM — Continuous Glucose Monitor. A wearable sensor that tracks blood glucose every few minutes. Signos members wear CGMs while eating meals they log.
● Related Foods: How celery's glucose response compares to other ultra-low-GI snack vegetables and the dips it lives with
Cucumber Glycemic Index & Glucose Score
Cucumber sits in the same near-zero-GL band as celery — useful for benchmarking what "noise-floor" vegetable response looks like in the snack context.
Carrots Glycemic Index & Glucose Score
Carrots are the other crudité staple — a slightly higher published GI than celery but a similar pattern of being defined by the dip alongside.
Bell Pepper Glycemic Index & Glucose Score
Bell pepper rounds out the crudité plate with the same low-GI, dip-dependent behavior — a good comparison point for raw-vegetable snack response.
Peanut Butter Glycemic Index & Glucose Score
Peanut butter is the fat-and-protein anchor that pulls celery's peak below baseline — see how peanut butter behaves on its own and across other pairings.
Hummus Glycemic Index & Glucose Score
Hummus is the other crudité dip that lands in the low-spike range — useful for understanding which dip choices keep raw-vegetable snacks clean.
Track Your Celery Response with Signos →
See how your glucose reacts to celery with peanut butter vs. in tuna salad vs. in soup — CGM data, personalized to you.

Frequently Asked Questions

Methodology

This page draws on 19,889 celery-containing logs from 5,021 unique Signos members, collected through the Signos production CGM platform. Data were filtered to postprandial glucose responses (PPGR) between 0 and 100 mg/dL to exclude sensor artifacts and implausible readings. Modifier effects were estimated using Welch's t-test on matched pairs against a defined baseline; all reported effects carry p<0.001 unless noted. We report mean glucose peak across a right-skewed distribution — the median spike of +22 mg/dL is a few mg/dL lower than the mean of +26.3 mg/dL, which is typical for this type of data. The food was identified in member logs using a regex matching "celery" as a primary ingredient, excluding celery seed and celery salt (seasoning, no carb signal). Any food-pair pairing slot requires at least 100 matched meals before it appears in the recipe builder.

Limitations

  • Self-reported portion sizes introduce measurement error — a logged "celery" may mean a thin garnish stalk or a full cup of diced mirepoix, affecting per-serving estimates.
  • Cohort skews health-motivated; Signos members actively tracking glucose tend to eat more mindfully than the general population, so real-world average spikes may differ.
  • The celery-root / celeriac sub-cohort is very small (n=50) and carries low confidence — its reported delta of +0.9 mg/dL vs. raw stalks should be treated as directional only and may shift with a larger sample.
  • The diced-cooked sub-cohort (n=367) carries medium confidence; the +4.2 mg/dL delta vs. raw stalks likely reflects the larger soup-and-bread meal context, not the cooking method itself.
  • Modifier analyses reflect meal-level correlations, not controlled feeding trials — causality cannot be established from observational CGM data alone.
  • Because celery's own glycemic load is 0, all food-pair deltas are functionally measurements of the paired food (peanut butter, ranch, soup-thickener, mayo in tuna salad) — not of the celery itself.
● Get your own data

See your own celery response

Across 19,889 logged meals containing celery, the average glucose peak was +26.3 mg/dL — but the IQR spans 23 mg/dL, meaning your personal response could land well above or below that. What you pair celery with matters too: peanut butter pulls the peak −2 mg/dL below baseline, ranch adds +2 mg/dL, and soup adds +5 mg/dL. A CGM shows you exactly where you land and which builds work best for your biology.

Start with Signos
50,000+
Members Measuring