Non-interactive mode

Non-interactive mode bypasses the TUI and prints results directly to stdout. Use it in CI pipelines, shell scripts, or whenever you want to process results programmatically.

Enabling non-interactive mode

Three equivalent methods:

1. Set the standard CI environment variable:

   CI=true github-code-search "useFeatureFlag" --org fulll

2. Use the explicit flag:

   github-code-search "useFeatureFlag" --org fulll --no-interactive

3. GitHub Actions — CI=true is set automatically:

   - run: github-code-search "useFeatureFlag" --org fulll
     env:
       GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}

TIP

In any environment where CI=true is set (GitHub Actions, GitLab CI, CircleCI, etc.), non-interactive mode is activated automatically — you don't need --no-interactive.

Example output

$ CI=true github-code-search "useFeatureFlag" --org fulll

3 repos · 5 files selected

- **fulll/auth-service** (2 matches)
  - [ ] [src/middlewares/featureFlags.ts:2:19](https://github.com/fulll/auth-service/blob/main/src/middlewares/featureFlags.ts#L2)
  - [ ] [tests/unit/featureFlags.test.ts:1:8](https://github.com/fulll/auth-service/blob/main/tests/unit/featureFlags.test.ts#L1)
- **fulll/billing-api** (2 matches)
  - [ ] [src/flags.ts:3:14](https://github.com/fulll/billing-api/blob/main/src/flags.ts#L3)
  - [ ] [src/routes/invoices.ts:1:1](https://github.com/fulll/billing-api/blob/main/src/routes/invoices.ts#L1)
- **fulll/frontend-app** (1 match)
  - [ ] [src/hooks/useFeatureFlag.ts:1:1](https://github.com/fulll/frontend-app/blob/main/src/hooks/useFeatureFlag.ts#L1)

replay command

github-code-search "useFeatureFlag" --org fulll --no-interactive

Replay command

At the end of every interactive session, github-code-search prints a replay command that encodes your exact selections as --exclude-repositories and --exclude-extracts arguments. Copy it to rerun the same filtered search in non-interactive mode:

github-code-search "useFeatureFlag" --org fulll --no-interactive \
  --exclude-repositories legacy-monolith \
  --exclude-extracts auth-service:tests/unit/featureFlags.test.ts:0

This is the recommended bridge between an interactive exploration session and a reproducible CI step.

GitHub Actions example

- name: Search for feature flag usages
  run: |
    github-code-search "useFeatureFlag" --org fulll \
      --format json \
      --output-type repo-only \
      > results.json
  env:
    GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}

Combining with shell tools

The Markdown output can be piped directly into other tools:

# Count how many repos contain the pattern (JSON is cleaner for scripting)
CI=true github-code-search "TODO" --org fulll \
  --format json --output-type repo-only \
  | jq '.selection.repos'

# Process JSON output with jq
CI=true github-code-search "useFeatureFlag" --org fulll --format json \
  | jq '[.results[].repo]'

What's next?

• Output formats — choose between Markdown and JSON
• Filtering — pre-exclude repos or extracts before running