The study was undertaken to evaluate the growth performance and feed utilization of African catfish, Clarias gariepinus, fed six diets (D) in which fishmeal (FM) was gradually replaced by a mixture of local plant by-products. In diets 1 and 2, FM (250 g kg−1) was replaced by sunflower oil cake (SFOC). In diets 3 and 4, FM (250 and 150 g kg−1, respectively) was replaced by SFOC and bean meal (BM) while FM was totally substituted by a mixture of groundnut oil cake (GOC), BM and SFOC in diets 5 and 6. Sunflower oil cake was cooked, soaked or dehulled in order to determine the appropriate processing techniques for improving the SFOC nutritive value and to evaluate the apparent digestibility coefficient (ADC) values of the alternative diets. No significant differences were observed for daily feed intake, weight gain, specific growth rate (SGR) and feed efficiency (FE) among fish fed D1, D2, D3 (250 g kg−1 FM), D4 (150 g kg−1 FM) and D6 (0 g kg−1 FM). The highest SGR (3.2% per day) and FE (1.2) were achieved in fish fed D3, and the lowest in fish fed D5 (0% FM), suggesting a maximum acceptable dietary concentration of hulled SFOC below 250 g kg−1 in African catfish juveniles. Protein efficiency ratio ranged from 2.2 to 3.2 for all dietary treatments and was positively influenced by FM inclusion. African catfish were able to digest plant protein very efficiently in all diets tested. ADC of protein ranged from 88.6 to 89.5%, while ADC of energy was relatively low for diets containing hulled sunflower oilcake (71–74%) and high when sunflower oilcake was dehulled (78.6–81.3%). Similarly, ADC of dry matter was higher when sunflower was dehulled (72.1%) when compared with crude SFOC (60.5%). Soaking increased ADC values for neutral detergent fibre (NDF), dry matter, energy, protein and amino acids (AA). There were no significant differences in protein ADCs (88–90%) with increased levels of dietary vegetable ingredients. Both soaking and dehulling of sunflower before incorporation helped in the reduction of NDF, antitrypsin and tannins. Digestibility of all AA was generally high, greater than 90% for both indispensable and non-indispensable AA. Based on the data obtained, it was possible to totally replace menhaden fish meal with a mixture of vegetable proteins (72% of total dietary protein) when diets contained a relatively low percentage of animal protein (28% based on blood meal and chicken viscera meal) without negative effects.