RAMSES is a new high-resolution N-body and hydrodynamical code designed to study structure formation in the universe with high spatial resolution. It uses Adaptive Mesh Refinement (AMR) with a tree-based data structure for recursive grid refinements. The N-body solver is similar to the ART code, while the hydrodynamical solver uses a second-order Godunov method for accurate thermal history computation. The code's accuracy is validated through various test cases, including pure gas dynamics and cosmological simulations. A refinement strategy is described, and potential spurious effects from shock wave propagation in the AMR grid are discussed. Results from a large simulation of a low-density ΛCDM universe with 256³ particles and 4.1×10⁷ cells in the AMR grid are reported, achieving a formal resolution of 8192³. Convergence analysis shows that numerical results converge to the code's resolution limit and match recent analytical predictions. The code uses a tree-based AMR structure, with a "Fully Threaded Tree" data structure. The N-body solver is inspired by the ART code, and the hydrodynamical solver is a second-order Godunov scheme. The code is tested on various scenarios, including acceleration around a point mass and particles in a ΛCDM simulation, demonstrating its accuracy and efficiency. The code's AMR implementation allows for high dynamic range and efficient computation, with time step control based on stability constraints. The refinement strategy ensures accurate treatment of physical problems while minimizing computational resources. The code is used in cosmological contexts with conformal time and super-comoving coordinates for comoving variables. The code's performance is validated through tests showing accurate acceleration calculations and convergence to analytical predictions.RAMSES is a new high-resolution N-body and hydrodynamical code designed to study structure formation in the universe with high spatial resolution. It uses Adaptive Mesh Refinement (AMR) with a tree-based data structure for recursive grid refinements. The N-body solver is similar to the ART code, while the hydrodynamical solver uses a second-order Godunov method for accurate thermal history computation. The code's accuracy is validated through various test cases, including pure gas dynamics and cosmological simulations. A refinement strategy is described, and potential spurious effects from shock wave propagation in the AMR grid are discussed. Results from a large simulation of a low-density ΛCDM universe with 256³ particles and 4.1×10⁷ cells in the AMR grid are reported, achieving a formal resolution of 8192³. Convergence analysis shows that numerical results converge to the code's resolution limit and match recent analytical predictions. The code uses a tree-based AMR structure, with a "Fully Threaded Tree" data structure. The N-body solver is inspired by the ART code, and the hydrodynamical solver is a second-order Godunov scheme. The code is tested on various scenarios, including acceleration around a point mass and particles in a ΛCDM simulation, demonstrating its accuracy and efficiency. The code's AMR implementation allows for high dynamic range and efficient computation, with time step control based on stability constraints. The refinement strategy ensures accurate treatment of physical problems while minimizing computational resources. The code is used in cosmological contexts with conformal time and super-comoving coordinates for comoving variables. The code's performance is validated through tests showing accurate acceleration calculations and convergence to analytical predictions.