GitHub

Cut the noise.
Own the filtering rules.

CCP, Command Compression Proxy, is a CLI output filtering layer for noisy terminal output before it reaches your coding agent.

Keep command behavior intact, shape terminal output with command-aware YAML rules, preserve exit codes and critical diagnostics, and fall back to native output when details matter. 

curl -fsSL https://raw.githubusercontent.com/SuppieRK/ccp/main/scripts/install.sh | sh

Terminal output is built for humans, not agents

Tools print banners, progress bars, repeated status lines, and other boilerplate that helps a person scan a terminal but adds little value for a coding agent reading shell output as context.

Noise wastes context and hides diagnostics

Long output burns the context window, slows follow-up reasoning, and makes failures, file paths, and critical diagnostics harder to spot when the agent decides what to do next.

CCP filters CLI output before agents see it

It removes repetitive shell noise before the output reaches the agent, while keeping the lines that still change what the agent should do next during terminal output filtering.

You own the rules and the behavior

Author YAML filters close to your workflow, preserve exit codes and critical diagnostics, and fall back to native output when it matters.

Use with your coding agent

Claude Code Codex OpenCode Cursor GitHub Copilot Gemini Windsurf Roo Code KiloCode Qwen Cline

Gains on real work

CCP does not just promise smaller output. Command ccp gain shows the token impact of real commands, so you can see where filtering helps, where it does not, and whether the output stays worth trusting.

Claude Code Gradle project 
88 cmds · 5,330,571 → 90,127 tokens (98.3% saved)
Wins : find (4.8m / 99%) · gradle (367k / 87%) · grep (6k / 1%)Drag : cd (23 cmds) · jar (21 cmds) · grep (4 cmds)Trend: ↑ +12.4 pts week over week (85.9% → 98.3%) · on a roll
Codex this repository 
1,825 cmds · 2,461,959 → 2,160,427 tokens (12.3% saved)
Wins : grep (67k / 60%) · go (54k / 90%) · git (16k / 59%)Drag : sed (765 cmds) · openspec (245 cmds)Trend: ↓ -2.1 pts week over week (14.4% → 12.3%) · slipping

Create CLI output filters for your workflow

Write command-aware YAML, replay real output, and verify exactly what changed before you promote a terminal output filter.

start with the command 
version: 1
filter: "yarn"
cases:
  - id: "run-success"
    when_arguments:
      have_sequence: ["run"]
    compress_output:
      stdout:
        lines:
          skip:
            - starts_with: "yarn run v"
            - starts_with: "$ "
            - starts_with: "Done in "
native output 
yarn run v1.22.22
$ node scripts/success.js
success-line-1
success-line-2
Done in 0.06s.

Scaffold a filter, capture a real command, verify the replay, then promote only when the behavior earns trust.

Read more about command-aware YAML filters.

Recorded examples from benchmark fixtures

The command on the left is the recorded fixture input. The right side is the verified CCP output for that same case.

npm run success-noisy

107 tokens -> 45 tokens

original 
> ccp-npm-benchmark-basic@1.0.0 success-noisy
> node --test test/success.test.js

TAP version 13
# Subtest: add computes sums
ok 1 - add computes sums
  ---
  duration_ms: 0.366266
  type: 'test'
  ...
# Subtest: divide computes quotient
ok 2 - divide computes quotient
  ---
  duration_ms: 0.067165
  type: 'test'
  ...
1..2
# tests 2
# suites 0
# pass 2
# fail 0
# cancelled 0
# skipped 0
# todo 0
# duration_ms 37.905034
compressed 
# Subtest: add computes sums
ok 1 - add computes sums
# Subtest: divide computes quotient
ok 2 - divide computes quotient
1..2
# tests 2
# pass 2
# fail 0
# duration_ms 37.905034