Soybeans can provide ruminally degradable protein, lipid, and metabolizable amino acid (AA) to lactating dairy cows; however, soy-based trypsin inhibitors can limit protein digestion in nonruminants. Eight ruminally cannulated lactating Holstein cows were used to evaluate the impacts of soy-based trypsin inhibitors on nutrient disappearance, lactation, and plasma AA bioavailability. Treatments were abomasal infusion of 0 or 400 g/d casein or a crystalline AA analog of casein with unroasted or roasted soybeans fed at 10% dry matter (DM). Data were analyzed using the MIXED procedure of SAS. Measures of digestion were determined from fecal output determined with acid detergent insoluble ash and urine output determined from measures of urine creatinine. Neither soybean processing (P ≥ 0.20) nor the source of abomasal infusion (P ≥ 0.60) impacted nutrient digestibility. Ruminal ammonia, isobutyrate, and isovalerate were increased (P ≤ 0.01) among cattle consuming unroasted soybeans. Source of infusion did not affect (P ≥ 0.38) ruminal volatile fatty acids or nitrogen metabolism. Ruminal N metabolism was largely unaffected by soybean processing although microbial N efficiency was greater (P < 0.01) among cows fed unroasted soybeans. DM intake and energy-corrected milk were greater (P < 0.01) in cows fed roasted compared to unroasted soybeans. The proportion of fat, protein, lactose, and solids not fat (SNF) in milk did not differ between soybean processing or postruminal AA source, but fat, protein, lactose, and SNF yield was greater (P ≤ 0.01) when cows were fed roasted soybeans because milk yields were greater when cows were fed roasted vs. unroasted soybeans. As expected, infusion of casein or its crystalline AA analog increased plasma essential AA and milk urea nitrogen concentration. The rate of increase in essential AA concentration in plasma was 2.9× greater for casein than for crystalline AA.