What do the states mean?+
- Solid: well studied, and it works.
- Half-there: the science is still forming.
- A ghost: almost no real studies.
- Hollow: well studied, and it does not work.
- A cord: the pair itself has been tested together.
- Air: nobody has ever tested this pair.
The full scoring method: How we score
Probiotics benefits: where the science stands
For gut and digestion: yes.
Gut and digestion are what most people take probiotics for, and the strongest evidence says it helps (a combined review covering 82 trials). As of August 2026, the verdicts on this page rest on evidence from 167 trials, drawn from the strongest research among the 23,972 papers on PubMed that mention probiotics. Every claim is checked below.
23,972 papers on PubMed · evidence from 167 trials
Claim by claim
- Gut and digestionWell studied, and it works. The biggest review covered 82 trials.
- Energy and fatigueThe science is still forming. Results lean positive, but there is not much yet.
- MoodWell studied, and it works. The biggest review covered 72 trials.
Pairs people ask about
The three sources we read
- 2023Meta-analysis · pools 82 trials, 10,332 people · small benefit for gut and digestionsource
- 2024Meta-analysis · pools 13 trials, 513 people · small benefit for energy and fatiguesource
- 2025Meta-analysis · pools 72 trials, 6,097 people · small benefit for moodsource
Chosen as the strongest evidence out of 23,972 papers on PubMed. Every row links to its paper.
Fair questions
Does probiotics work?+
For gut and digestion, yes. The biggest review covered 82 trials.
Is probiotics safe?+
No caution is recorded on this page. That is not a safety guarantee; the studies we read were not safety trials.
Why only 3 sources, when PubMed lists 23,972?+
The 23,972 is every paper that mentions probiotics in any way: animal studies, blood measurements, papers where it is a side note. The strongest sources are combined reviews, and each one already gathers every trial its authors could find. Our 3 cover 167 trials between them, and every one is linked above. Both counts are measured, not estimated.