We evaluated the effects of tannins from Flemingia macrophylla (CIAT 17403) and Calliandra calothyrsus (San Ramón CIAT 22310 and Patulul CIAT 22316) on in vitro ruminal and post-ruminal dry matter and apparent protein degradation. For each tannin source (legumes), different dosages of polyethylene glycol (PEG) (8000 Da) in McDougall buffer were added to achieve ratios of 0:3, 1:3, 2:3 and 3:3 PEG:condensed tannin (CT). Ruminal fluid mixed with McDougall buffer (1:4) was added to tubes containing only legume foliage (control) or PEG-treated legume foliage. For both Calliandra varieties, a higher ruminal dry matter degradation was observed at a PEG:CT ratio of 3:3. For F. macrophylla, no differences were found between 2:3 and 3:3 ratios (p > 0.05), indicating that a PEG:CT ratio of 2:3 might be enough to bind tannins. Increasing PEG:CT ratios increased apparent ruminal degraded protein and ammonia concentration (p < 0.0001) differing among species (species × ratio: p < 0.0001). The degradation of bypass crude protein (dBCP) was influenced by both legume type and PEG:CT ratio (p < 0.0001). For Patulul, as PEG:CT ratio increased, dBCP increased, but after tannin ratio of 2:3, there was not a significant increase, and for San Ramón, dBCP degradation was higher as PEG:CT ratio increased up to 2:3. For Flemingia, dBCP was higher than PEG:CT ratio of 0:3 but not different among 1:3, 2:3 or 3:3. Low concentration of CT (116 mg/g DM) increased the proportion of protein digested in the abomasum, but higher levels of CT (252 mg/g) clearly reduced the proportion of digested CP. For Flemingia, PEG:CT ratio of 2:3 is enough to inactivate tannins, while PEG:CT ratio of 3:3 was needed for Calliandra and consequently increased ruminal degradation of dry mater (rdDM), and crude protein (rdCP), total degradation of dry matter (tdDM), crude protein (tdCP) and ammonia levels.