Complementary feeding practices are often inadequate in developing countries, resulting in a significant nutritional decline between 6 and 18 mo of age. We assessed the effectiveness of an educational intervention to promote adequate complementary feeding practices that would be feasible to sustain with existing resources. The study was a cluster randomized controlled trial in communities in the state of Haryana in India. We developed the intervention through formative research. Eight communities were pair matched on their baseline characteristics; one of each pair was randomly assigned to receive the intervention and the other to no specific feeding intervention. Health and nutrition workers in the intervention communities were trained to counsel on locally developed feeding recommendations. Newborns were enrolled in all of the communities (552 in the intervention and 473 in the control) and followed up every 3 mo to the age of 18 mo. The main outcome measures were weights and lengths at 6, 9, 12, and 18 mo and complementary feeding practices at 9 and 18 mo. All analyses were by intent to treat. In the overall analyses, there was a small but significant effect on length gain in the intervention group (difference in means 0.32 cm, 95% CI, 0.03, 0.61). The effect was greater in the subgroup of male infants (difference in mean length gain 0.51 cm, 95% CI 0.03, 0.98). Weight gain was not affected. Energy intakes from complementary foods overall were significantly higher in the intervention group children at 9 mo (mean ± SD: 1556 ± 1109 vs. 1025 ± 866 kJ; P < 0.001) and 18 mo (3807 ± 1527 vs. 2577 ± 1058 kJ; P < 0.001). Improving complementary feeding practices through existing services is feasible but the effect on physical growth is limited. Factors that limit physical growth in such settings must be better understood to plan more effective nutrition programs.