diff --git a/docs/source/algorithms.md b/docs/source/algorithms.md
index bd8837b9a..54c7e8ea6 100644
--- a/docs/source/algorithms.md
+++ b/docs/source/algorithms.md
@@ -3920,48 +3920,36 @@ addition to optimagic when using an NLOPT algorithm. To install nlopt run
 
 optimagic supports the [IMINUIT MIGRAD Optimizer](https://iminuit.readthedocs.io/). To
 use MIGRAD, you need to have
-[the iminuit package](https://github.com/scikit-hep/iminuit) installed (pip install
-iminuit).
+[the iminuit package](https://github.com/scikit-hep/iminuit) installed
+(`pip install iminuit`).
 
 ```{eval-rst}
 .. dropdown::  iminuit_migrad
 
-    .. code-block::
-
-        "iminuit_migrad"
-
-    `MIGRAD <https://iminuit.readthedocs.io/en/stable/reference.html#iminuit.Minuit.migrad>`_ is 
-    the workhorse algorithm of the MINUIT optimization suite, which has been widely used in the 
-    high-energy physics community since 1975. The IMINUIT package is a Python interface to the 
-    Minuit2 C++ library developed by CERN.
+    **How to use this algorithm:**
 
-    Migrad uses a quasi-Newton method, updating the Hessian matrix iteratively
-    to guide the optimization. The algorithm adapts dynamically to challenging landscapes
-    using several key techniques:
+    .. code-block::
 
-    - **Quasi-Newton updates**: The Hessian is updated iteratively rather than recalculated at 
-      each step, improving efficiency.
-    - **Steepest descent fallback**: When the Hessian update fails, Migrad falls back to steepest
-      descent with line search.
-    - **Box constraints handling**: Parameters with bounds are transformed internally to ensure 
-      they remain within allowed limits.
-    - **Heuristics for numerical stability**: Special cases such as flat gradients or singular 
-      Hessians are managed using pre-defined heuristics.
-    - **Stopping criteria based on Estimated Distance to Minimum (EDM)**: The optimization halts 
-      when the predicted improvement becomes sufficiently small.
-              
-    For details see :cite:`JAMES1975343`.
+        import optimagic as om
+        om.minimize(
+          ...,
+          algorithm=om.algos.iminuit_migrad(stopping_maxfun=10_000, ...)
+        )
+        
+    or
+        
+    .. code-block::
 
-    **Optimizer Parameters:**  
+        om.minimize(
+          ...,
+          algorithm="iminuit_migrad",
+          algo_options={"stopping_maxfun=10_000, ...}
+        )
 
-    - **stopping.maxfun** (int): Maximum number of function evaluations. If reached, the optimization stops 
-      but this is not counted as successful convergence. Function evaluations used for numerical gradient 
-      calculations do not count toward this limit. Default is 1,000,000.
+    **Description and available options:**
 
-    - **n_restarts** (int): Number of times to restart the optimizer if convergence is not reached.
+    .. autoclass:: optimagic.optimizers.iminuit_migrad.IminuitMigrad
 
-      - A value of 1 (the default) indicates that the optimizer will only run once, disabling the restart feature.  
-      - Values greater than 1 specify the maximum number of restart attempts.  
 ```
 
 ## Nevergrad Optimizers
diff --git a/src/optimagic/optimizers/iminuit_migrad.py b/src/optimagic/optimizers/iminuit_migrad.py
index 4c6490e74..27721131b 100644
--- a/src/optimagic/optimizers/iminuit_migrad.py
+++ b/src/optimagic/optimizers/iminuit_migrad.py
@@ -42,8 +42,44 @@
 )
 @dataclass(frozen=True)
 class IminuitMigrad(Algorithm):
+    """Minimize a scalar differentiable function using the MIGRAD algorithm from
+    iminuit.
+
+    This optimizer wraps the MIGRAD algorithm from the iminuit package, which provides a
+    Python interface to the Minuit2 C++ library developed and maintained by CERN.
+
+    MIGRAD is a local optimization method in the quasi-Newton family. It iteratively
+    builds an approximation of the inverse Hessian matrix using the DFP variable-metric
+    method to efficiently navigate optimization landscapes.
+
+    At each iteration, the algorithm attempts a Newton step, using gradient and Hessian
+    approximations to move toward the function’s minimum. If this step fails to reduce
+    the objective function, MIGRAD conducts a line search along the gradient direction
+    to maintain progress. This continues until the convergence criteria, such as the
+    Estimated Distance to Minimum (EDM) are met, that is, they fall below preset
+    thresholds.
+
+    MIGRAD is designed for statistical optimization problems where accurate parameter
+    uncertainty estimates are essential. It excels at maximum-likelihood and least-
+    squares fits common in scientific computing, and is best suited for smooth,
+    differentiable cost functions.
+
+    For best performance, supply analytical gradients. Convergence and solution will
+    depend on your starting values. Bound constraints (limits) supported.
+
+    """
+
     stopping_maxfun: int = STOPPING_MAXFUN
+    """Maximum number of function evaluations."""
+
     n_restarts: int = N_RESTARTS
+    """Number of times to restart the optimizer if convergence is not reached.
+
+    A value of 1 (the default) indicates that the optimizer will only run once,
+    disabling the restart feature. Values greater than 1 specify the maximum number of
+    restart attempts.
+
+    """
 
     def _solve_internal_problem(
         self, problem: InternalOptimizationProblem, params: NDArray[np.float64]