diff --git a/README.md b/README.md
index 0c80e08..081f97a 100644
--- a/README.md
+++ b/README.md
@@ -2,9 +2,9 @@
 [![Docker Stars](https://img.shields.io/docker/stars/sjdonado/bit.svg)](https://hub.docker.com/r/sjdonado/bit)
 [![Docker Image Size](https://img.shields.io/docker/image-size/sjdonado/bit/latest)](https://hub.docker.com/r/sjdonado/bit)
 
-Lightweight URL shortener (API-only) with minimal resource requirements. Avg memory consumption under pressure is around **60MiB**, avg CPU load 60%.
+Lightweight URL shortener (API-only) with minimal resource requirements. Avg memory consumption **30MiB**, avg CPU load 20%.
 
-Highly performant: **6.6k req/sec**, latency 18.9ms (100k requests using 125 connections, [benchmark](docs/SETUP.md#benchmark)).
+Highly performant: **7.9k req/sec**, latency 15.8ms (100k requests using 125 connections, [benchmark](docs/SETUP.md#benchmark)).
 
 Self-hosted: [Dokku](docs/SETUP.md#dokku), [Docker Compose](docs/SETUP.md#docker-compose).
 
@@ -18,7 +18,7 @@ It is feature-complete by design: simple and reliable without unnecessary bloat.
 - Multiple users are supported via API key authentication. Users can create, list and delete keys via the [CLI](docs/SETUP.md#cli).
 - Easy to extend, Ruby on Rails-inspired setup.
 
-## Minimum Requirements
+## Recommented requirements
 - 100MB disk space
 - 70MiB RAM
 - x86_64 or ARM64
diff --git a/app/controllers/click.cr b/app/controllers/click.cr
index cee343d..b855510 100644
--- a/app/controllers/click.cr
+++ b/app/controllers/click.cr
@@ -14,7 +14,6 @@ module App::Controllers
 
     @@processor_started = begin
       spawn do
-        batch_size = 64
         batch = [] of NamedTuple(
           link_id: Int64,
           remote_address: String,
@@ -22,21 +21,30 @@ module App::Controllers
           referer: String
         )
 
+        batch_size = 32
+        min_batch = 32
+        max_batch = 512
+        scaling_factor = 1.5
+
         loop do
           select
           when click_data = @@click_channel.receive
             batch << click_data
 
-            # Collect clicks until we have a batch or a timeout
             if batch.size >= batch_size
               process_click_batch(batch)
               batch.clear
+              # Increase batch size when reaching capacity
+              batch_size = (batch_size * scaling_factor).to_i.clamp(min_batch, max_batch)
             end
-          when timeout(0.5.seconds)
-            # Process whatever we have after timeout
+          when timeout(1.seconds)
             unless batch.empty?
               process_click_batch(batch)
               batch.clear
+              batch_size = (batch_size / scaling_factor).to_i.clamp(min_batch, max_batch)
+            else
+              # Reset to default if idle
+              batch_size = min_batch unless batch_size == min_batch
             end
           end
         end
diff --git a/benchmark.cr b/benchmark.cr
index 18ac1b1..aa46bcd 100755
--- a/benchmark.cr
+++ b/benchmark.cr
@@ -176,7 +176,6 @@ def analyze_resource_usage
   puts "Analyzing resource usage..."
 
   sleep 2.seconds
-  # Read stats directly from file for more accurate results
   if File.exists?(STATS_FILE)
     lines = File.read_lines(STATS_FILE)
     # Skip header
diff --git a/docs/SETUP.md b/docs/SETUP.md
index 1df014c..8ef5b03 100644
--- a/docs/SETUP.md
+++ b/docs/SETUP.md
@@ -126,43 +126,42 @@ Waiting for the application to be ready...
 Seeding the database...
 Checking seed results...
 Fetching all created links from /api/links...
-Selected link for benchmarking: http://localhost:4000/slug4280
+Selected link for benchmarking: http://localhost:4000/slug2576
 Starting benchmark with Bombardier...
-Bombarding http://localhost:4000/slug4280 with 100000 request(s) using 125 connection(s)
- 100000 / 100000 [==============================================================] 100.00% 6562/s 15s
+Bombarding http://localhost:4000/slug2576 with 100000 request(s) using 125 connection(s)
+ 100000 / 100000 [===================================================================================================================================================================] 100.00% 7795/s 12s
 Done!
 Statistics        Avg      Stdev        Max
-  Reqs/sec      6609.73    1508.34   13145.76
-  Latency       18.92ms     2.34ms    74.58ms
+  Reqs/sec      7900.70    7570.15   29263.59
+  Latency       15.89ms    10.22ms    67.32ms
   Latency Distribution
-     50%    18.83ms
-     75%    20.19ms
-     90%    21.80ms
-     95%    23.10ms
-     99%    26.54ms
+     50%     4.82ms
+     75%     9.24ms
+     90%    51.61ms
+     95%    52.74ms
+     99%    55.07ms
   HTTP codes:
     1xx - 0, 2xx - 0, 3xx - 100000, 4xx - 0, 5xx - 0
     others - 0
-  Throughput:     1.80MB/s
+  Throughput:     2.14MB/s
 Benchmark completed successfully.
 Analyzing resource usage...
 Timestamp       CPU(%)  Memory(MiB)
-1742732843      0.02    44.71
-1742732845      0.02    44.71
-1742732847      85.34   69.55
-1742732849      83.5    69.93
-1742732851      84.26   69.97
-1742732853      83.64   70.01
-1742732855      84.23   70.04
-1742732857      86.41   69.17
-1742732859      85.77   69.2
-1742732861      59.67   68.55
+1742763202      0.01    44.83
+1742763204      0.01    44.78
+1742763206      91.53   68.23
+1742763208      92.03   68.17
+1742763210      91.0    68.09
+1742763212      92.73   68.38
+1742763214      92.17   67.66
+1742763216      91.1    67.69
+1742763218      2.93    67.04
 
 **** Resource Usage Statistics ****
-  Measurements: 10
-  Average CPU Usage: 65.29%
-  Average Memory Usage: 64.58 MiB
-  Peak CPU Usage: 86.41%
-  Peak Memory Usage: 70.04 MiB
+  Measurements: 9
+  Average CPU Usage: 61.5%
+  Average Memory Usage: 62.76 MiB
+  Peak CPU Usage: 92.73%
+  Peak Memory Usage: 68.38 MiB
 Cleanup completed. Resource usage data saved in resource_usage.txt
 ```