This study aimed to evaluate the use of wattle tannin as an adsorbent of toxic metals in broilers fed on diets with different levels of replacement of limestone with golden mussel meal. First, we conducted a trial to evaluate the performance of broiler chicks from 1 to 21 days old when fed on wattle tannin. For this we used 720 male broiler chicks that were one day old, in a completely randomized design with six treatments and six replications. The treatments were composed of basal diets with increasing levels of wattle tannin (0, 250, 500, 750, 1000 and 1250 g tonne-1), where the tannin replaced the inert material of the feed. All of the diets were isonutritive and isocaloric. The results showed that 250 g tonne-1 of tannin was detrimental to weight gain, final weight and feed conversion, and the use of 1000 g tonne-1 of wattle tannin, in addition to these effects, caused a drop in consumption of the ration. There was no effect of treatment on morphometric measurements of villi, crypt or the villus:crypt ratio. Next, we conducted a second test to evaluate the performance of broiler chicks from 1 to 21 days of age when fed with increasing levels of replacement (0, 25, 50, 75, 100%) of limestone with golden mussel and even supplementation or not wattle tannin. Five repetitions were used and the birds were housed in a completely randomized design. In diets with the inclusion of tannin, 250 g tonne-1 of wattle tannin was used, which replaced the inert material in the feed. In this test, 1200 male broiler chicks that were one day old were used. The results showed that the golden mussel contains large quantities of calcium, with low concentrations of toxic metals and microbiological contamination within the allowed range. Performance data show that mussel can be used to replace up to 100% of limestone in feed without affecting the variables, but the use of 250 g tonne-1 of wattle tannin affects the performance of chicks. Bone quality was not changed by the treatments. Phosphorus and calcium in the bones remained stable during treatment and only the gray variation showed interaction effects from the use of mussel and wattle tannin.