-mi verbs are a class of athematic verbs in the Proto-Indo-European language. The name derives from the first person singular form of the verbs in the present indicative active in Ancient Greek, which is the dictionary form. The person marker is -mi, e.g. ειμί (eimí, I am), δίδωμι (dídōmi, I give), φημί (phēmí, I say) etc. The conjugation of mi-verbs differs from the much more common thematic, or "omega" (-ω, -ō), conjugation. Mi-verbs were "regularised" (transformed to their -o counterparts, or replaced altogether) in the Koine era and the transformation was almost complete by Byzantine times, with some vestiges of the -mi conjugation surviving only in the Passive voice in Modern Greek.
Mi-verbs are an extremely ancient feature of Proto-Indoeuropean grammar. Sanskrit verbs are exclusively -mi verbs, all Latin verbs except the form "sum" of "esse" are -o verbs, whereas Ancient Greek verbs can be either. The typical example used to illustrate this is the Sanskrit verb bhárami, Latin fero, Greek φέρω, phérō, (I bring).
The most common verb is the copula, and therefore it has survived longest, just like in Latin the only Anglo-Saxon mi-verb is the copula "eom", nowadays shortened to "am". The verb "give" is also very common and is together with the copula still a mi-verb in modern Slavic languages like the Czech "dávám".