Alvarezsauridae is a family of small, long-legged dinosaurs. Although originally thought to represent the earliest known flightless birds, they are now thought to be an early diverging branch of maniraptoran theropods. Alvarezsaurids were highly specialized. They had tiny but stout forelimbs, with compact, bird-like hands. Their skeletons suggest that they had massive breast and arm muscles, possibly adapted for digging or tearing. They had long, tube-shaped snouts filled with tiny teeth. They have been interpreted as myrmecophagous, adapted to prey on colonial insects such as termites, with the short arms acting as effective digging instruments to break into nests. This has however been contested.

Alvarezsaurus, the type genus of the family, was named after the historian and physician Gregorio Álvarez.

Bonaparte (1991) described the first alvarezsaurid, Alvarezsaurus calvoi, from an incomplete skeleton found in Patagonia, Argentina. Bonaparte also named a family, Alvarezsauridae, to contain it. He argued that Alvarezsaurus might be most closely related to the ornithomimosaurs.

In 1993, Perle et al. described the next alvarezsaur to be discovered, naming it Mononychus olecranus (meaning "one claw"). A month later they changed the genus name to Mononykus, because the earlier spelling was already the genus name of an extant beetle. Perle et al. mistakenly described Mononykus as a member of Avialae, one more advanced than Archaeopteryx. They argued that the family Alvarezsauridae was actually a group of Mesozoic flightless birds on the basis of several features that were supposedly unique to birds.

In 1996, Novas described another member of the group called Patagonykus puertai. Karhu and Rautian (1996) described a Mongolian member of the family; Parvicursor remotus. Chiappe et al.(1998) described another Mongolian member, Shuvuuia deserti, and found it to be a bird as in Perle et al.'s analysis.

These mistaken assignments of alvarezsaurids to birds were caused primarily by features that are strikingly, or even uniquely, avian. The sternum, for example, is elongated and deeply keeled for an enlarged pectoralis muscle, as it is in neognathous birds and volant ratites. One bone in the skull of Shuvuuia appeared to be an ectethmoid fused to a prefrontal. The ectethmoid is an ossification known only in Neornithes. Other birdlike characters included the palatine, foramen magnum, cervical and caudal vertebrae, and many others.

Several researchers disagreed with Perle et al. (1993) and Chiappe et al. (1998), Feduccia (1994), Ostrom (1994), Wellnhofer (1994), Kurochkin (1995), Zhou (1995), and Sereno (1997) considered it unlikely that alvarezsaurids were members of Avialae. Martin (1997) performed a cladistic analysis but Sereno criticized it strongly, finding it flawed by incorrect codings, use of only select data, and results that did not support his conclusions. Sereno (1999) performed a new analysis, revising the anatomical interpretations and clarifying the characters. He found that alvarezsaurids were more parsimoniously related to the Ornithomimosauria.

As the more primitive members of the Alvarezsauridae were better characterized, the monophyly of the clade was strongly supported, but the more primitive members lacked the most birdlike traits. Some of these traits had been misinterpreted as well. The remaining similarities between birds and alvarezsaurs—like the keeled sterna—are another case of homoplasy, where the derived alvarezsaurids developed birdlike characters through convergent evolution, rather than inheriting them from a common ancestor with birds.