About This Desk

What this desk is

Restorative Peptides is a literature digest covering three research peptides studied for tissue repair: BPC-157, TB-500 and the two-peptide pairing known as Wolverine. It is independent. It sells nothing, dispenses nothing, treats no one and takes no position on whether anybody should use any compound described here.

It exists because of a specific gap. The tissue-repair peptide literature is discussed almost entirely in terms of whether something works, while the question people actually arrive with is how long — how long before anything changes, how long an effect lasts, how long a study followed its subjects before deciding it had found something. Those are answerable questions, and answering them honestly turns out to be the most useful thing a digest in this field can do, because the answers are frequently 'nobody measured that.'

Every page here is built from a single set of numbered sources, listed once on the references page and referred to by number throughout. There is no separate reading list per page and no source cited anywhere that is not in that list.

How the record is read

Four editorial rules govern every page.

The species is stated with the interval. A benefit measurable from day 14 in a rat ^[9] is written as a rat result, not as a general fact about healing. The tissue-repair peptide literature is overwhelmingly animal work — a 2025 systematic review found 35 of its 36 included studies were preclinical ^[13] — and a timeline assembled from rodent studies is a rodent timeline. Rescaling it into a human schedule is the single most common error in this subject, and this desk does not make it.

Anecdote is labelled and kept separate. Research-use communities have produced a large body of dated recovery accounts. They are reported here because they are part of how these compounds are actually discussed, they are always marked anecdotal rather than clinical, and they never carry doses.

Absence is reported as a finding. When no study establishes an onset, a duration or a persistence window, the page says so rather than filling the space with a plausible-sounding estimate. On the Wolverine page that absence is most of the page.

The molecule studied is distinguished from the molecule sold. Where the encouraging research used full-length thymosin beta-4 rather than the short fragment marketed as TB-500 ^[10], the difference is stated at the point of use, not buried in a caveat at the bottom.

What this desk will not do

It does not recommend, suggest or imply a dose for any person. Where a dose appears — 2, 12 or 18 mg/kg in a rat stroke study ^[9], or a 14-day intravenous escalation in healthy volunteers ^[11] — it is reported exactly as the study reported it, with the species and the design attached, because that is what makes the finding legible. It is never a schedule.

It does not give medical advice, evaluate whether a finding applies to any individual, or comment on any specific product, supplier or vendor. None of the three compounds covered here is an approved medicine; all are sold for laboratory research use only, and all are prohibited in tested sport. A 2025 narrative review concluded that BPC-157 should be treated as investigational because rigorous large-scale human trials are lacking ^[2], and a 2026 review of the wider class found favourable animal-model outcomes alongside scarce human safety data, potential for serious harm and operation largely outside regulatory oversight ^[8]. Those verdicts belong in an 'about' page as much as anywhere, because they define the limits of everything else on the site.

Corrections and currency

This is a literature digest, and a literature moves. Reviews are superseded, trials report, and a compound's regulatory status can change without any new science behind it.

A correction that improves the accuracy of the record is the most valuable message this desk receives, and the contact page explains what makes one easy to act on. Where a claim rests on a single study, the page says so. Where it rests on a review rather than primary data, the page says that too.