Oviraptor (/ˈoʊvɪræptər/; lit. 'egg thief') is a genus of oviraptorid dinosaur that lived in Asia during the Late Cretaceous period. The first remains were collected from the Djadokhta Formation of Mongolia in 1923 during a paleontological expedition led by Roy Chapman Andrews, and in the following year the genus and type species Oviraptor philoceratops were named by Henry Fairfield Osborn. The genus name refers to the initial thought of egg-stealing habits, and the specific name was intended to reinforce this view indicating a preference over ceratopsian eggs. Despite the fact that numerous specimens have been referred to the genus, Oviraptor is only known from a single partial skeleton regarded as the holotype, as well as a nest of about fifteen eggs and several small fragments from a juvenile.

Oviraptor was a rather small feathered oviraptorid, estimated at 1.6–2 m (5.2–6.6 ft) long with a weight between 33–40 kg (73–88 lb). It had a wide lower jaw with a skull that likely had a crest. Both upper and lower jaws were toothless and developed a horny beak, which was used during feeding along the robust morphology of the lower jaws. The arms were well-developed and elongated ending in three fingers with curved claws. Like other oviraptorids, Oviraptor had long hindlimbs that had four-toed feet, with the first toe reduced. The tail was likely not very elongated, and ended in a pygostyle that supported large feathers.

The initial relationships of Oviraptor were poorly understood at the time and was assigned to the unrelated Ornithomimidae by the original describer, Henry Osborn. However, re-examinations made by Rinchen Barsbold proved that Oviraptor was distinct enough to warrant a separate family, the Oviraptoridae. When first described, Oviraptor was interpreted as an egg-thief, egg-eating dinosaur given the close association of the holotype with a dinosaur nest. However, findings of numerous oviraptorosaurs in nesting poses have demonstrated that this specimen was actually brooding the nest and not stealing nor feeding on the eggs. Moreover, the discovery of remains of a small juvenile or nestling have been reported in association with the holotype specimen, further supporting parental care.

The first remains of Oviraptor were discovered on reddish sandstones of the Late Cretaceous Djadokhta Formation of Mongolia, at the Bayn Dzak locality (also known as Flaming Cliffs), during the Third Central Asiatic expedition in 1923. This expedition was led by the North American naturalist Roy Chapman Andrews and ended in the discovery of three new-to-science theropod fossil remains—including those of Oviraptor. These were formally described by the North American paleontologist Henry Fairfield Osborn in 1924, who in the basis of the new material, named the genera Oviraptor, Saurornithoides and Velociraptor. The particular genus Oviraptor was erected with the type species O. philoceratops based on the holotype AMNH 6517, a partial individual lacking the back of the skeleton but including a badly crushed skull, partial cervical and dorsal vertebrae, pectoral elements including the furcula with the left arm and partial hands, the left ilium and some ribs. Accordingly, this specimen was found lying over a nest of approximately 15 eggs—a nest that has been catalogued as AMNH 6508—with the skull separated from the eggs by only 10 cm (100 mm) of sediment. Given the close proximity of both specimens, Osborn interpreted Oviraptor as a dinosaur with egg-eating habits, and explained that the generic name, Oviraptor, is Latin for "egg seizer" or "egg thief", due to the association of the fossils. The specific name, philoceratops, is intended as "fondness for ceratopsian eggs" which is also given as a result of the initial thought of the nest pertaining to Protoceratops or another ceratopsian. However, Osborn suggested that the name Oviraptor could reflect an incorrect perception of this dinosaur. Furthermore, Osborn found Oviraptor to be similar to the unrelated—at the time, however, considered related—fast-running ornithomimids based on the toothless jaws, and assigned Oviraptor to the Ornithomimidae. Osborn had previously reported the taxon as "Fenestrosaurus philoceratops", but this was later discredited.

In 1976, the Mongolian paleontologist Rinchen Barsbold noted some inconsistencies regarding the taxonomic placement of Oviraptor and concluded that this taxon was quite distinct from ornithomimids based on anatomical traits. Under this consideration, he erected the Oviraptoridae to contain Oviraptor and close relatives. After Osborn's initial description of Oviraptor, the egg nest associated with the holotype was accepted to have belonged to Protoceratops, and oviraptorids were largely considered to have been egg-eating theropods. Nevertheless, in the 1990s, the discovery of numerous nesting and nestling oviraptorid specimens proved that Osborn was correct in his caution regarding the name of Oviraptor. These findings showed that oviraptorids brooded and protected their nests by crouching on them. This new line of evidence showed that the nest associated with the holotype of Oviraptor belonged to it and the specimen was actually brooding the eggs at the time of death, not preying on them.

After the naming of Oviraptoridae in 1976, Barsbold referred six additional specimens to Oviraptor, including two particular specimens under the number MPC-D 100/20 and 100/21. In 1986, Barsbold realized that the latter two did not belong to the genus and instead they represented a new oviraptorid: Conchoraptor. Most of the other specimens are also unlikely to belong to Oviraptor itself, and they have been assigned to other oviraptorids. A partial individual also with eggs from the Bayan Mandahu Formation of China was referred in 1996 by Dong Zhiming and Philip J. Currie, the specimen IVPP V9608. However, in 2010 Nicholas R. Longrich and the two latter paleontologist have expressed their uncertainties regarding this referral as there are several anatomical differences such as the hand phalangeal proportions. They concluded that this specimen was a different and indeterminate species not referrable to this taxon. In 1981, Barsbold referred the specimen MPC-D 100/42 to Oviraptor, a very well-preserved and rather complete individual from the Djadokhta Formation. Since the known elements of Oviraptor were so fragmentary compared to other members, MPC-D 100/42 became the prime reference/depiction of this taxon being prominently labelled as Oviraptor philoceratops in scientific literature.

This conception was refuted by James M. Clark and colleagues in 2002, who noted that this tall-crested specimen has more features of the skull in common with Citipati than it does with Oviraptor—which in fact, does not preserve a crest—and it may represent a second species of the former, or, an entire new genus. In 1986, Barsbold described a second species of Oviraptor, "O. mongoliensis", based on specimen MPC-D 100/32a which hails from the Nemegt Formation. However, a re-examination by Barsbold in 1997 found enough differences in this specimen to name the new genus Rinchenia, but he did not describe it with formality and this new oviraptorid remained as a nomen dubium. This was amended by the Polish paleontologist Halszka Osmólska and team in 2004 by formally naming the taxon Rinchenia mongoliensis. The North American paleontologist Mark A. Norell and colleagues in 2018 reported a new specimen of Oviraptor: AMNH 33092, which is composed of a tibia and two metatarsals of a nestling or very small juvenile. AMNH 33092 was found in association with the holotype and it was likely part of the nest. Oviraptor is now known from the holotype with associated eggs, and a juvenile/nestling.

The holotype specimen has been estimated at 1.6–2 m (5.2–6.6 ft) in length with a weight ranging from 33 to 40 kg (73 to 88 lb). Though the holotype largely lacks the posterior region of the skeleton, it is likely that Oviraptor had two well-developed hindlimbs that ended in three functional toes with the first one being vestigial, as well as a relatively reduced tail. As evidenced in related oviraptorids, the arms were covered by elongated feathers, and the tail ended in a pygostyle, which is known to support a fan of feathers.