OBJECTIVE: To formulate age- and context-specific complementary feeding recommendations (CFR) for infants and young children (IYC) and to compare the potential of filling population-level nutrient gaps using common sets of CFR across age groups. DESIGN: Linear programming was used to develop CFR using locally available and acceptable foods based on livelihood- and age-group-specific dietary patterns observed through 24 h dietary recalls. Within each livelihood group, the nutrient potential of age-group-specific v. consolidated CFR across the three age groups was tested. SETTING: Three food-insecure counties in northern Kenya; namely, settled communities from Isiolo (n 300), pastoralist communities from Marsabit (n 283) and agro-pastoralist communities from Turkana (n 299). SUBJECTS: Breast-fed IYC aged 6-23 months (n 882). RESULTS: Age-specific CFR could achieve adequacy for seven to nine of eleven modelled micronutrients, except among 12-23-month-old children in agro-pastoralist communities. Contribution of Fe, Zn and niacin remained low for most groups, and thiamin, vitamin B6 and folate for some groups. Age-group-consolidated CFR could not reach the same level of nutrient adequacy as age-specific sets among the settled and pastoralist communities. CONCLUSIONS: Context- and age-specific CFR could ensure adequate levels of more modelled nutrients among settled and pastoralist IYC than among agro-pastoralist communities where use of nutrient-dense foods was limited. Adequacy of all eleven modelled micronutrients was not achievable and additional approaches to ensure adequate diets are required. Consolidated messages should be easier to implement as part of a behaviour change strategy; however, they would likely not achieve the same improvements in population-level dietary adequacy as age-specific CFR.