diff --git a/lib/credo/check.ex b/lib/credo/check.ex
index 64297f384..023b5f0ab 100644
--- a/lib/credo/check.ex
+++ b/lib/credo/check.ex
@@ -393,8 +393,11 @@ defmodule Credo.Check do
 
       defp do_run_on_all_source_files(exec, source_files, params) do
         source_files
-        |> Enum.map(&Task.async(fn -> run_on_source_file(exec, &1, params) end))
-        |> Enum.each(&Task.await(&1, :infinity))
+        |> Task.async_stream(fn source -> run_on_source_file(exec, source, params) end,
+          max_concurrency: exec.max_concurrent_check_runs,
+          timeout: :infinity
+        )
+        |> Stream.run()
 
         :ok
       end
diff --git a/lib/credo/check/runner.ex b/lib/credo/check/runner.ex
index e6c3574c8..99f3bd794 100644
--- a/lib/credo/check/runner.ex
+++ b/lib/credo/check/runner.ex
@@ -19,9 +19,14 @@ defmodule Credo.Check.Runner do
       |> warn_about_ineffective_patterns(exec)
       |> fix_deprecated_notation_for_checks_without_params()
 
-    Credo.Check.Worker.run(check_tuples, exec.max_concurrent_check_runs, fn check_tuple ->
-      run_check(exec, check_tuple)
-    end)
+    check_tuples
+    |> Task.async_stream(
+      fn check_tuple ->
+        run_check(exec, check_tuple)
+      end,
+      timeout: :infinity
+    )
+    |> Stream.run()
 
     :ok
   end
diff --git a/lib/credo/check/worker.ex b/lib/credo/check/worker.ex
deleted file mode 100644
index 1652e85e3..000000000
--- a/lib/credo/check/worker.ex
+++ /dev/null
@@ -1,126 +0,0 @@
-defmodule Credo.Check.Worker do
-  @moduledoc false
-
-  @doc """
-  Runs all members of `workloads` using ``.
-  """
-  def run(workloads, max_concurrency, work_fn) do
-    {:ok, server_pid} = GenServer.start_link(__MODULE__.Server, workloads)
-
-    worker_context = %{
-      runner_pid: self(),
-      server_pid: server_pid,
-      max_concurrency: max_concurrency,
-      work_fn: work_fn,
-      results: []
-    }
-
-    outer_loop(worker_context, 0)
-  end
-
-  @doc """
-  Called when a workload has finished.
-  """
-  def send_workload_finished_to_runner(worker_context, _workload, result) do
-    send(worker_context.runner_pid, {self(), {:workload_finished, result}})
-  end
-
-  defp outer_loop(worker_context, taken) do
-    available = worker_context.max_concurrency - taken
-
-    cond do
-      available <= 0 ->
-        wait_for_workload_finished(worker_context, taken)
-
-      taken_workloads = __MODULE__.Server.take_workloads(worker_context.server_pid, available) ->
-        inner_loop(worker_context, taken_workloads, taken)
-
-      # we fall thru here if there are no checks left
-      # there are two options: we are done ...
-      taken == 0 ->
-        {:ok, worker_context.results}
-
-      # ... or we need for the very last batch to finish up
-      true ->
-        wait_for_workload_finished(worker_context, taken)
-    end
-  end
-
-  defp wait_for_workload_finished(worker_context, taken) do
-    receive do
-      {_spawned_pid, {:workload_finished, result}} ->
-        # IO.puts("Finished #{workload}")
-        new_worker_context = %{worker_context | results: [result | worker_context.results]}
-
-        outer_loop(new_worker_context, taken - 1)
-    end
-  end
-
-  defp inner_loop(worker_context, [], taken) do
-    outer_loop(worker_context, taken)
-  end
-
-  defp inner_loop(worker_context, [workload | rest], taken) do
-    spawn_fn = fn ->
-      result = worker_context.work_fn.(workload)
-
-      send_workload_finished_to_runner(worker_context, workload, result)
-    end
-
-    spawn_link(spawn_fn)
-
-    inner_loop(worker_context, rest, taken + 1)
-  end
-
-  defmodule Server do
-    @moduledoc false
-    @timeout :infinity
-
-    use GenServer
-
-    def take_workloads(pid, count) do
-      GenServer.call(pid, {:take_workloads, count}, @timeout)
-    end
-
-    #
-    # Server
-    #
-
-    @impl true
-    def init(workloads) do
-      state = %{
-        waiting: nil,
-        workloads: workloads
-      }
-
-      {:ok, state}
-    end
-
-    @impl true
-    def handle_call({:take_workloads, count}, from, %{waiting: nil} = state) do
-      {:noreply, take_workloads(%{state | waiting: {from, count}})}
-    end
-
-    defp take_workloads(%{waiting: nil} = state) do
-      state
-    end
-
-    defp take_workloads(%{waiting: {from, _count}, workloads: []} = state) do
-      GenServer.reply(from, nil)
-
-      %{state | waiting: nil}
-    end
-
-    defp take_workloads(%{workloads: []} = state) do
-      state
-    end
-
-    defp take_workloads(%{waiting: {from, count}, workloads: workloads} = state) do
-      {reply, workloads} = Enum.split(workloads, count)
-
-      GenServer.reply(from, reply)
-
-      %{state | workloads: workloads, waiting: nil}
-    end
-  end
-end
diff --git a/lib/credo/execution.ex b/lib/credo/execution.ex
index 506ddcf0f..005d62d42 100644
--- a/lib/credo/execution.ex
+++ b/lib/credo/execution.ex
@@ -42,13 +42,15 @@ defmodule Credo.Execution do
             enable_disabled_checks: nil,
             ignore_checks_tags: [],
             ignore_checks: nil,
-            max_concurrent_check_runs: nil,
             min_priority: 0,
             mute_exit_status: false,
             only_checks_tags: [],
             only_checks: nil,
             read_from_stdin: false,
 
+            # This is no longer used, but we keep it so existing plugins that use it don't break
+            max_concurrent_check_runs: nil,
+
             # state, which is accessed and changed over the course of Credo's execution
             pipeline_map: %{},
             commands: %{},