Cell’s ability to deform, migrate or sense their mechanical environment is of prime importance at all stages of an organism’s life from embryonic development, to physiological and pathological situations. The main actor of these properties is the actin cytoskeleton, an assembly of organized bio-polymer structures whose properties are finely regulated by actin binding proteins (ABPs). AFM or micropipettes have been used to probe the mechanical properties of actin networks as well as their assembly in presence of opposing forces. However these techniques are limited in the number of measurements that can be achieved. We developed a new approach based on magnetic colloids, spherical or lab-made cylindrical, that allows quantitative high-throughput experiments to be carried out. We focus here on Arp2/3 actin which are implicated in a variety of cellular functions such as motility or endocytosis. We use inter-dipolar forces that develop between superparamagnetic particles in presence of magnetic field to apply controlled stresses to actin networks assembled either from purified mammalian proteins or from yeast extracts. The purified system only requires a minimal set of proteins: branching protein (Arp2/3), capping protein (gelsolin) and depolymerizing factor (ADF/cofilin) and gives a very good control on the concentration of the proteins. The advantage of the yeast extracts approach is that proteins can be genetically removed one-by-one, in order to test for their functions in a near-physiological environment. The self-organization of our particles in a magnetic field allows many actin networks to be probed at the same time improving the throughput of mechanical measurements. The magnetic micro-cylinders for which we specifically developed a soft-lithography based fabrication procedure allow the assembly process in a presence of opposing forces to be precisely followed as well as non linear measurements to be carried out directly. More specifically I will show here that the mechanics of branched networks –in the absence of crosslinker- is non linear and seems to vanish when there is no stress pointing to an entanglements-based mechanics. In parallel, I will show our first results about the dependence of the growth on opposing stresses. I will also discuss mechanical properties of yeast actin networks by comparing purified systems and extracts either from the wild type or from strain lacking given crosslinkers. These actin networks exhibit some plasticity and the absence of the different crosslinkers have different impacts on the mechanical properties.