diff --git a/searches/astar.py b/searches/astar.py
new file mode 100644
index 000000000000..5c1de087f6aa
--- /dev/null
+++ b/searches/astar.py
@@ -0,0 +1,146 @@
+#!/usr/bin/env python3
+
+"""
+Pure Python implementation of the A* (A-star) pathfinding algorithm.
+
+For doctests run:
+    python3 -m doctest -v astar.py
+
+For manual testing run:
+    python3 astar.py
+"""
+
+import heapq
+from typing import Callable, Iterable, Tuple, List, Dict, Optional, Set
+
+# Type aliases for readability
+Node = Tuple[int, int]
+NeighborsFn = Callable[[Node], Iterable[Tuple[Node, float]]]
+HeuristicFn = Callable[[Node, Node], float]
+
+
+def astar(
+    start: Node, goal: Node, neighbors: NeighborsFn, h: HeuristicFn
+) -> Optional[List[Node]]:
+    """
+    A* algorithm for pathfinding on a graph defined by a neighbor function.
+
+    A* maintains:
+      -> g[n]: cost from start to node n (best known so far)
+      -> f[n] = g[n] + h(n, goal): estimated total cost of a path through n to goal
+      -> open_list: min-heap of candidate nodes prioritized by smallest f-score
+      -> closed_list: set of nodes already expanded (best path to them fixed)
+
+    :param start: starting node
+    :param goal: target node
+    :param neighbors: function returning (neighbor, step_cost) pairs for a node
+    :param h: admissible heuristic h(n, goal) estimating remaining cost
+    :return: list of nodes from start to goal (inclusive), or None if no path
+
+    Examples:
+    >>> def _h(a, b):  # Manhattan distance
+    ...     (x1, y1), (x2, y2) = a, b
+    ...     return abs(x1 - x2) + abs(y1 - y2)
+    >>> def _nbrs(p):  # 4-connected grid, unit costs, unbounded
+    ...     x, y = p
+    ...     return [((x + 1, y), 1), ((x - 1, y), 1), ((x, y + 1), 1), ((x, y - 1), 1)]
+    >>> astar((0, 0), (2, 2), _nbrs, _h)[-1]
+    (2, 2)
+    """
+    # Min-heap of (f_score, node). We only store (priority, node) to keep it simple;
+    # if your nodes aren't directly comparable, add a tiebreaker counter to the tuple.
+    open_list: List[Tuple[float, Node]] = []
+
+    # Nodes we've fully explored (their best path is finalized).
+    closed_list: Set[Node] = set()
+
+    # g-scores: best known cost to reach each node from start
+    g: Dict[Node, float] = {start: 0.0}
+
+    # Parent map to reconstruct the path once we reach the goal
+    parent: Dict[Node, Optional[Node]] = {start: None}
+
+    # Initialize the frontier with the start node (f = h(start, goal))
+    heapq.heappush(open_list, (h(start, goal), start))
+
+    while open_list:
+        # Pop the node with the smallest f-score (best promising path so far)
+        _, current = heapq.heappop(open_list)
+
+        # If we've already expanded this node via a better path, skip it
+        if current in closed_list:
+            continue
+        closed_list.add(current)
+
+        # Goal check: reconstruct the path by following parents backward
+        if current == goal:
+            path: List[Node] = []
+            while current is not None:
+                path.append(current)
+                current = parent[current]
+            return path[::-1]  # reverse to (start ... goal)
+
+        # Explore current's neighbors
+        for neighbor, cost in neighbors(current):
+            # If neighbor was already finalized, ignore
+            if neighbor in closed_list:
+                continue
+
+            # Tentative g-score via current
+            tentative_g = g[current] + cost
+
+            # If this is the first time we see neighbor, or we found a cheaper path to it
+            if neighbor not in g or tentative_g < g[neighbor]:
+                g[neighbor] = tentative_g
+                parent[neighbor] = current
+                f_score = tentative_g + h(neighbor, goal)
+                heapq.heappush(open_list, (f_score, neighbor))
+
+    # If the frontier empties without reaching the goal, no path exists
+    return None
+
+
+def heuristic(n: Node, goal: Node) -> float:
+    """
+    Manhattan (L1) distance heuristic for 4-connected grid movement with unit costs.
+    Admissible and consistent for axis-aligned moves.
+
+    :param n: current node
+    :param goal: target node
+    :return: |x1 - x2| + |y1 - y2|
+    """
+    x1, y1 = n
+    x2, y2 = goal
+    return abs(x1 - x2) + abs(y1 - y2)
+
+
+def neighbors(node: Node) -> Iterable[Tuple[Node, float]]:
+    """
+    4-neighborhood on an unbounded grid with unit edge costs.
+
+    Replace/extend this for:
+      -> bounded grids (check bounds before yielding)
+      -> obstacles (skip blocked cells)
+      -> diagonal moves (add the 4 diagonals with cost sqrt(2) and switch heuristic)
+
+    :param node: (x, y) coordinate
+    :return: iterable of ((nx, ny), step_cost)
+    """
+    x, y = node
+    return [
+        ((x + 1, y), 1),
+        ((x - 1, y), 1),
+        ((x, y + 1), 1),
+        ((x, y - 1), 1),
+    ]
+
+
+if __name__ == "__main__":
+    # Example usage / manual test
+    start: Node = (0, 0)
+    goal: Node = (5, 5)
+    path = astar(start, goal, neighbors, heuristic)
+    print("Path found:", path)
+    # Expected (one optimal path; yours may differ but length should be 10 moves + start):
+    # Path found: [(0, 0), (0, 1), (0, 2), (0, 3), (0, 4), (0, 5),
+    #              (1, 5), (2, 5), (3, 5), (4, 5), (5, 5)]