The effect of different condensed tannins concentrations on protein metabolism from browse supplements was investigated in a 90-day trial using sixty-six male Ethiopian Menz sheep. Teff straw (Eragrostis tef) was fed ad libitum (control diet), or supplemented with 190 g dried leaves of six Sesbania sesban accessions with increasing levels of condensed tannin (S1, S2, S3, S4, S5 and S6), lablab (Dolichos lablab), tagasaste (Chamaecytisus palmensis), leucaena (Leucaena leucocephala) and goetzei (Sesbania goetzei) in a completely randomized block design. The condensed tannins were highest in goetzei; S6, S5 and leucaena, intermediate in S4, S3, S2 and S1, and lowest in tagasaste and lablab. The supplemented animals had significantly (P < 0.05) higher total dry matter and nitrogen (N) intake than the ones fed teff straw alone. The digestibility of N was lower for the control diet than for any other treatment (P < 0.05). S1 and S2 supplemented diets had significantly higher (P < 0.05) N digestibilities than all other diets. Faecal N, urinary N and urinary N per kg N excreted were significantly different (P < 0.0001) between diets. With increasing tannin levels (among Sesbania accessions) there was a significant decrease (P < 0.05) in urinary N (S1 > S2 > S3 > S4 > S5), and an increase (P < 0.05) in faecal N (S1 < S2 < S3 < S4 < S5). Supplementation increased faecal N output significantly (P < 0.0001) as well as the N retention. Among the forage supplements, N retention was significantly (P < 0.0001) lower in lablab-, tagasaste-, leucaena-, S4- and goetzei-supplemented diets, than for S1, S2, S5 and S6. Apparent nitrogen digestibility was positively correlated (P < 0.001) with the supplement dry matter and crude protein (CP) degradation after 24 h (r = 0.93 and r = 0.85, respectively), the CP content (r = 0.87), and was negatively correlated with acid detergent fibre and neutral detergent fibre (r = –0.87 and –0.87, respectively). The CP degradability characteristics of the forages differed (P < 0.001) in water solubility (93–470 g kg-1 CP), rate of degradation (2.58–9.73 %/h), lag phase (–1.36–13.37 h), and estimated escape protein (262–619 g kg-1 CP). With increasing tannin levels (among Sesbania accessions), there was a significant decrease (P < 0.0001) in the rate of degradation (S1 > S2 > S3 > S4 > S5), and an increase in the estimated escape protein. The estimated rumen degradable protein (supplements) varied from 482 to 744 g kg-1 CP, while intestine digestible protein and the undegradable protein varied from 140 to 314 g kg-1 CP, hence the browses can supply adequate levels of rumen degradable and bypass protein.