The diagnosis of visceral leishmaniasis (VL) in humans and animal reservoir hosts is difficult, particularly in rural areas where the disease is endemic and laboratory facilities are limited. This study aimed to develop a latex agglutination test (LAT) for the rapid detection of anti-Leishmania antibodies against the A2 antigen derived from the amastigote form as well as those against crude antigens derived from the promastigote form of an Iranian strain of Leishmania (Leishmania) infantum. The A2 antigen (42-100 kDa) was prepared from the amastigote form of L. infantum, purified with electroelution and compared with the crude antigen from the promastigote form of L. infantum. Both antigens showed appropriate intensity reactions, were selected using dot blotting of positive and negative pooled sera and used to sensitize 0.9-μm latex beads. The tests were carried out on sera from 43 symptomatic, human patients with VL confirmed by parasitological examination and direct agglutination test (DAT), 30 healthy controls and 32 patients with other infections but without VL. Canine sera were collected from 63 domestic dogs with VL confirmed using parasitological examinations and DAT and 31 healthy dogs from areas non-endemic for VL. Compared with the controls, human sera from DAT-confirmed patients yielded a sensitivity of 88.4% (95% CI, 82.1-94.5%) and specificity of 93.5% (95% CI, 87.0-99.7%) on A2-LAT (amastigote) when 1:3200 was used as the cut-off titre. A good degree of agreement was found between A2-LAT and DAT (0.914). LAT required 3-5 min to complete, versus the 12-18 h needed for DAT. Compared with the controls, A2-LAT of canine sera from DAT-confirmed cases yielded a sensitivity of 95.2% (95% CI, 95.0-95.4%) and specificity of 100% (95% CI 100%) when 1:320 was used as the cut-off titre. A good degree of agreement was found between A2-LAT and DAT (0.968). Similarly, the sensitivity and specificity of Pro.-LAT (promastigote) was calculated to be 88.4% and 91.9%, respectively for human sera and 96.8% and 90.3%, respectively for canine sera. No statistically significant differences were observed between A2-LAT and Pro.-LAT for the detection of human and canine L. infantum infections. In conclusion, A2-LAT and Pro.-LAT could be used in parallel to screen for L. infantum infections in humans and dogs in areas endemic for VL in Iran.