diff --git a/.github/workflows/bencher.yml b/.github/workflows/bencher.yml
new file mode 100644
index 00000000..f35b22a8
--- /dev/null
+++ b/.github/workflows/bencher.yml
@@ -0,0 +1,32 @@
+on:
+  push:
+
+jobs:
+  benchmark_base_branch:
+    name: Continuous Benchmarking with Bencher
+    permissions:
+      contents: read
+      checks: write
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@v4
+      - uses: bencherdev/bencher@main
+      - name: Build release binary
+        run: cargo build --release
+      - name: Add fgfa to PATH
+        run: echo "$PWD/target/release" >> $GITHUB_PATH
+      - name: Fetch test data
+        run: make fetch
+
+        # FlatGFA Benchmarks 
+      - name: Track file size benchmarks with Bencher
+        run: |
+          bencher run \
+          --project flatgfa \
+          --token '${{ secrets.BENCHER_API_TOKEN }}' \
+          --branch main \
+          --testbed ubuntu-latest \
+          --err \
+          --adapter json \
+          --github-actions '${{ secrets.GITHUB_TOKEN }}' \
+          python bench/benchmark_web.py mini_bencher 10 del 
\ No newline at end of file
diff --git a/.github/workflows/build.yml b/.github/workflows/build.yml
index 56b9e154..2c68ee95 100644
--- a/.github/workflows/build.yml
+++ b/.github/workflows/build.yml
@@ -2,8 +2,6 @@ name: build
 
 on:
   push:
-    branches:
-      - main
   pull_request:
     branches:
       - main
diff --git a/bench/benchmark_web.py b/bench/benchmark_web.py
new file mode 100644
index 00000000..13920031
--- /dev/null
+++ b/bench/benchmark_web.py
@@ -0,0 +1,208 @@
+import sys
+import os
+import json
+import subprocess
+from pathlib import Path
+import time
+import tomllib
+import gzip
+import shutil
+
+# Parse the GFA URLs from graphs.toml
+with open("bench/graphs.toml", "rb") as f:
+    toml_graphs = tomllib.load(f)
+
+hprc_dict = dict(toml_graphs["hprc"])
+ 
+test_dict = dict(toml_graphs["test"]) 
+
+gont_dict = dict(toml_graphs["1000gont"])
+ 
+smoke_files = [test_dict["k"]]
+ 
+mini_files = [test_dict["lpa"], test_dict["chr6c4"], hprc_dict["chrM"]]
+
+med_files = [hprc_dict["chr20"], hprc_dict["chrX"], gont_dict["chr16"]]
+
+big_files = [hprc_dict["chrY"], hprc_dict["chr1"], hprc_dict["chr10"]]
+
+results = "filesize_benchmark.txt"
+
+# Download a GFA file from the internet
+def download_file(target_name, web_file):
+  gzipped = False
+  temp_name = ""
+  if "gfa.gz" in web_file:
+    gzipped = True
+  if gzipped:
+    temp_name = f"{target_name}.gz"
+
+  if not Path(target_name).exists():
+    if gzipped:
+      subprocess.run(["curl", "-o", temp_name, web_file],
+              check = True) 
+      with gzip.open(temp_name, "rb") as f_in:
+        with open(target_name, "wb") as f_out:
+          shutil.copyfileobj(f_in, f_out)
+      subprocess.run(["rm", "-rf", temp_name], check = True) 
+    else:
+      subprocess.run(["curl", "-o", target_name, web_file],
+              check = True) 
+
+# Run a single test      
+def test(command, test_file_name, num_iter):
+  if command == "extract":
+    with open(os.devnull, "w") as devnull:
+      start_time = time.time()
+      for _ in range(num_iter):
+        subprocess.run(["fgfa", "-I", test_file_name, "extract", "-n", "3", "-c", "3"], stdout=devnull,
+            stderr=devnull,
+            check=True) 
+      end_time = time.time()
+      return ((end_time - start_time) * 1000) / num_iter
+        
+  elif command == "chop":
+    with open(os.devnull, "w") as devnull:
+      start_time = time.time()
+      for _ in range(num_iter):
+        subprocess.run(["fgfa", "-I", test_file_name, "chop", "-c", "3", "-l"], stdout=devnull,
+            stderr=devnull,
+            check=True) 
+      end_time = time.time()
+      return ((end_time - start_time) * 1000) / num_iter
+
+  elif command == "depth":
+    with open(os.devnull, "w") as devnull:
+      start_time = time.time()
+      for _ in range(num_iter):
+        subprocess.run(["fgfa", "-I", test_file_name, "depth"], stdout=devnull,
+            stderr=devnull,
+            check=True) 
+      end_time = time.time()
+      return ((end_time - start_time) * 1000) / num_iter
+  return 0.0
+  
+# Run the latency benchmark across all test files
+def benchmark(test_config):
+  del_cond = ""
+  norm_cond = ""
+  num_iter = 0
+  iter_count = -1
+  
+  # Read command-line arguments
+  if len(sys.argv) >= 3:
+    iter_count = int(sys.argv[2]) # Can be any integer
+  
+  if len(sys.argv) >= 4:
+    del_cond = sys.argv[3] # Can be "del", "_", or not provided
+
+  if len(sys.argv) >= 5:
+    norm_cond = sys.argv[4] # Can be "norm", or not provided
+
+  # Choose test file set
+  test_files = []
+  if "smoke" in test_config:
+    test_files = smoke_files
+    num_iter = 2
+  elif "mini" in test_config:
+    test_files = mini_files
+    num_iter = 10
+  elif "med" in test_config:
+    test_files = med_files
+    num_iter = 5
+  elif "big" in test_config:
+    test_files = big_files
+    num_iter = 2
+  else:
+    raise ValueError("Incorrect test config provided")
+  
+  # Set number of test iterations
+  if not iter_count == -1:
+   num_iter = iter_count
+  
+  i = 0
+  total_time = 0.0
+  extract_time = 0.0
+  chop_time = 0.0
+  depth_time = 0.0
+  size_bytes_avg = 0
+
+  # Run a test for each file in the set
+  for file in test_files:
+    test_file_name = f"tests/{test_config}_{i}.gfa"
+    download_file(test_file_name, file)
+    subprocess.run(["fgfa", "-I", test_file_name, "-o", results],
+                  check = True) 
+    size_bytes_avg += os.path.getsize(results)
+    extract_time += test("extract", test_file_name, num_iter)
+    chop_time += test("chop", test_file_name, num_iter)
+    depth_time += test("depth", test_file_name, num_iter)
+    subprocess.run(["rm", "-rf", results], check = True) 
+
+    # Delete test files if flag set
+    if del_cond == "del":
+      subprocess.run(["rm", "-rf", test_file_name], check = True) 
+    i += 1
+  if (norm_cond == "norm"):
+
+    # Write new normalization values
+    with open("bench/normalization.toml", "w") as f:
+      f.write("[normalization_factors]\n")
+      f.write(f"extract = {extract_time}\n")
+      f.write(f"chop = {chop_time}\n")
+      f.write(f"depth = {depth_time}\n")
+    return (1.0, size_bytes_avg)
+  else:
+
+    # Read normalization values
+    with open("bench/normalization.toml", "rb") as f:
+      data = tomllib.load(f)
+    extract_norm = data["normalization_factors"]["extract"]
+    chop_norm = data["normalization_factors"]["chop"]
+    depth_norm = data["normalization_factors"]["depth"]
+
+    # Normalize values
+    extract_time /= extract_norm
+    chop_time /= chop_norm
+    depth_time /= depth_norm
+
+    # Return the harmonic mean
+    size_bytes_avg /= len(test_files)
+    return (3 / ((1 / extract_time) + (1 / chop_time) + (1 / depth_time)), size_bytes_avg / 1000.0)
+
+# Read the desired test file set from command-line input
+test_config = ""
+if len(sys.argv) >= 2:
+  test_config = sys.argv[1] # Can be either "smoke", "mini", "med", or "big"
+else:
+  raise ValueError("No arguments provided")
+
+bench_results = benchmark(test_config)
+          
+
+# Output the benchmark results, either in a Bencher JSON format, or a standard 
+# command-line format
+if "bencher" in test_config:
+  bencher_json = {
+    "FlatGFA Benchmark Results": {
+      "Average Execution Latency": {"value": round(bench_results[0], 2)}, 
+      "Average File Size": {"value": round(bench_results[1], 2)},
+    }
+  }
+  json.dump(bencher_json, sys.stdout)
+else:
+
+  # Only print latency info if flag set
+  if "latency" in test_config:
+    print(f"Average Execution Latency: {round(bench_results[0], 2)} ms")
+
+  # Only print filesize info if flag set
+  elif "filesize" in test_config:
+    print(f"Average File Size: {round(bench_results[1], 2)} KB")
+  else:
+    print(f"Average Execution Latency: {round(bench_results[0], 2)} ms")
+    print(f"Average File Size: {round(bench_results[1], 2)} KB")
+  
+
+# Command format: python bench/latency_benchmark_web.py [size](_bencher/_latency/_filesize) [run_count] (del/_) (norm)
+# () = optional, [] = replace with value  
\ No newline at end of file
diff --git a/bench/filesize_benchmark.py b/bench/filesize_benchmark.py
new file mode 100644
index 00000000..4d1733fc
--- /dev/null
+++ b/bench/filesize_benchmark.py
@@ -0,0 +1,32 @@
+import sys
+import os
+import json
+import subprocess
+import tomllib
+
+def benchmark(test_file):
+  subprocess.run(["fgfa", "-I", test_file, "-o", "filesize_benchmark.txt"], 
+                 check = True) 
+  size_bytes = os.path.getsize("filesize_benchmark.txt")
+  subprocess.run(["rm", "-rf", "filesize_benchmark.txt"], check = True) 
+  return size_bytes
+
+gfa_files = ["tests/chr6.C4.gfa", "tests/DRB1-3123.gfa", "tests/LPA.gfa"]
+sizes = {name: float(benchmark(name)) / 1000.0 for name in gfa_files}
+size_bytes_avg = (sizes["tests/chr6.C4.gfa"] + sizes["tests/DRB1-3123.gfa"] +
+                   sizes["tests/DRB1-3123.gfa"]) / 3.0
+
+bencher_json = {
+  "FlatGFA File Size": {
+    "chr6.C4 (File Size)": {"value": round(sizes["tests/chr6.C4.gfa"], 2)}, 
+    "DRB1-3123 (File Size)": {"value": round(sizes["tests/DRB1-3123.gfa"], 2)}, 
+    "LPA (File Size)": {"value": round(sizes["tests/DRB1-3123.gfa"], 2)}, 
+    "Average (File Size)": {"value": round(size_bytes_avg, 2)}
+  }
+}
+
+json.dump(bencher_json, sys.stdout)
+
+
+
+
diff --git a/bench/filesize_benchmark_web.py b/bench/filesize_benchmark_web.py
new file mode 100644
index 00000000..178e789d
--- /dev/null
+++ b/bench/filesize_benchmark_web.py
@@ -0,0 +1,109 @@
+import sys
+import os
+import json
+import subprocess
+from pathlib import Path
+import tomllib
+import gzip
+import shutil
+
+# Parse the GFA URLs from graphs.toml
+with open("bench/graphs.toml", "rb") as f:
+    toml_graphs = tomllib.load(f)
+
+hprc_dict = dict(toml_graphs["hprc"])
+ 
+test_dict = dict(toml_graphs["test"]) 
+
+gont_dict = dict(toml_graphs["1000gont"])
+
+smoke_files = [test_dict["k"]]
+ 
+mini_files = [test_dict["lpa"], test_dict["chr6c4"], hprc_dict["chrM"]]
+
+med_files = [hprc_dict["chr20"], hprc_dict["chrX"], gont_dict["chr16"]]
+ 
+big_files = [hprc_dict["chrY"], hprc_dict["chr1"], hprc_dict["chr10"]]
+
+results = "filesize_benchmark.txt"
+
+# Download a GFA file from the internet
+def download_file(target_name, web_file):
+  gzipped = False
+  temp_name = ""
+  if "gfa.gz" in web_file:
+    gzipped = True
+  if gzipped:
+    temp_name = f"{target_name}.gz"
+
+  if not Path(target_name).exists():
+    if gzipped:
+      subprocess.run(["curl", "-o", temp_name, web_file],
+              check = True) 
+      with gzip.open(temp_name, "rb") as f_in:
+        with open(target_name, "wb") as f_out:
+          shutil.copyfileobj(f_in, f_out)
+      subprocess.run(["rm", "-rf", temp_name], check = True) 
+    else:
+      subprocess.run(["curl", "-o", target_name, web_file],
+              check = True) 
+  
+# Run the file size benchmark across all files
+def benchmark(test_config):
+  test_cond = ""
+  if len(sys.argv) >= 3:
+    test_cond = sys.argv[2] # Can be "del", or not provided
+
+  # Choose test file set
+  test_files = []
+  if "smoke" in test_config:
+    test_files = smoke_files
+  elif "mini" in test_config:
+    test_files = mini_files
+  elif "med" in test_config:
+     test_files = med_files
+  elif "big" in test_config:
+    test_files = big_files
+  else:
+    raise ValueError("Incorrect test config provided")
+
+  size_bytes_avg = 0
+  i = 0
+
+  # Run a test for each file in the set
+  for file in test_files:
+    test_file_name = f"tests/{test_config}_{i}.gfa"
+    download_file(test_file_name, file)
+    subprocess.run(["fgfa", "-I", test_file_name, "-o", results],
+                  check = True) 
+    size_bytes = os.path.getsize(results)
+    subprocess.run(["rm", "-rf", results], check = True) 
+    if test_cond == "del":
+      subprocess.run(["rm", "-rf", test_file_name], check = True) 
+    size_bytes_avg += size_bytes
+    i += 1
+  size_bytes_avg /= len(test_files)
+  return size_bytes_avg / 1000.0 
+
+# Read the desired test file set from command-line input
+test_config = ""
+if len(sys.argv) >= 2:
+  test_config = sys.argv[1] # Can be either "smoke", "mini", "med", or "big"
+else:
+  raise ValueError("No arguments provided")
+
+
+# Output the benchmark results, either in a Bencher JSON format, or a standard 
+# command-line format
+if "bencher" in test_config:
+  bencher_json = {
+    "FlatGFA File Size Average": {
+      "Average File Size": {"value": round(benchmark(test_config), 2)}, 
+    }
+  }
+  json.dump(bencher_json, sys.stdout)
+else:
+  print(f"File Size Average: {round(benchmark(test_config), 2)} KB")
+
+# Command format: python bench/filesize_benchmark_web.py [size](_bencher) (del) 
+# () = optional, [] = replace with value
diff --git a/bench/latency_benchmark.py b/bench/latency_benchmark.py
new file mode 100644
index 00000000..fdb9c97b
--- /dev/null
+++ b/bench/latency_benchmark.py
@@ -0,0 +1,46 @@
+import sys
+import time
+import os
+import json
+import subprocess
+
+
+
+def benchmark(test_file, num_iter):
+  total_time = 0.0
+  for i in range(num_iter):
+    start_time = time.time()
+    with open(os.devnull, "w") as devnull:
+      subprocess.run(["fgfa", "-I", test_file, "extract", "-n", "3", "-c", "3"], stdout=devnull,
+          stderr=devnull,
+          check=True) 
+    end_time = time.time()
+    total_time += (end_time - start_time) * 1000
+  return total_time / num_iter
+
+avg_time = 0.0
+
+benchmark("tests/DRB1-3123.gfa", 1) # warmup rounds
+benchmark("tests/chr6.C4.gfa", 1)
+benchmark("tests/LPA.gfa", 1)
+time_1 = benchmark("tests/chr6.C4.gfa", 10)
+time_2 = benchmark("tests/DRB1-3123.gfa", 10)
+time_3 = benchmark("tests/LPA.gfa", 10)
+avg_time = (time_1 + time_2 + time_3) / 3
+
+
+
+bencher_json = {
+  "FlatGFA Extract Latency": {
+    "chr6.C4 (Running Time)": {"value": round(time_1, 2)}, 
+    "DRB1-3123 (Running Time)": {"value": round(time_2, 2)}, 
+    "LPA (Running Time)": {"value": round(time_3, 2)}, 
+    "Average (Running Time)": {"value": round(avg_time, 2)}
+  }
+}
+
+json.dump(bencher_json, sys.stdout)
+
+
+
+
diff --git a/bench/normalization.toml b/bench/normalization.toml
new file mode 100644
index 00000000..fb8e864f
--- /dev/null
+++ b/bench/normalization.toml
@@ -0,0 +1,4 @@
+[normalization_factors]
+extract = 34.96956825256348
+chop = 288.8672351837158
+depth = 15.636205673217773
diff --git a/normalization.toml b/normalization.toml
new file mode 100644
index 00000000..94557d13
--- /dev/null
+++ b/normalization.toml
@@ -0,0 +1,4 @@
+[normalization_factors]
+extract = 40.43102264404297
+chop = 292.3729419708252
+depth = 15.982627868652344
diff --git a/process.py b/process.py
index 3d5c89ed..07b0deba 100644
--- a/process.py
+++ b/process.py
@@ -17,7 +17,7 @@ def format_json_data(node_depths, mem="segments0"):
     depths = node_depths["memories"][mem]
     print("#node.id\tdepth")
     for i in range(len(depths)):
-        print(f"{i+1}\t{depths[i]}")
+        print(f"{i + 1}\t{depths[i]}")
 
 
 if __name__ == "__main__":
diff --git a/tests/.gitignore b/tests/.gitignore
index 6533033c..4a323ae8 100644
--- a/tests/.gitignore
+++ b/tests/.gitignore
@@ -27,4 +27,4 @@ subset-paths/*.out
 
 depth/*.out
 depth/basic/*.out
-depth/subset-paths/*.out
+depth/subset-paths/*.out
\ No newline at end of file