filemerger-cli 0.3.2

Developer CLI tool to consolidate project files into a single output

FileMerger

FileMerger is a developer-focused CLI tool that consolidates project files into a single plain-text output.

It is designed to help developers:

• Share complete code context with AI tools (ChatGPT, Gemini, Grok, Claude, etc.)
• Review large codebases
• Create audit or snapshot files
• Prepare structured input for analysis

---

Installation

pip install filemerger-cli

---

Basic Usage

Merge a directory:

filemerger src/

Specify output file:

filemerger src/ --output context.txt

Dry run (no file written):

filemerger . --dry-run

---

Output Modes

FileMerger supports multiple output modes depending on who (or what) will consume the output.

1. Default Mode (Human-Readable)

filemerger src/

Use this when:

• You want to read the output yourself
• You are reviewing or auditing code
• You want clear visual separation

Characteristics:

• File lists and headers
• Visual separators
• Structured, readable layout

---

2. LLM Mode (--llm)

filemerger src/ --llm

Use this when:

• The output will be pasted into an AI system
• You want deterministic file references
• You want to reduce semantic noise

Characteristics:

• Files are numbered ([1], [2], …)
• No decorative separators
• Simple, predictable structure

Example:

[1] path/to/file.py
<content>

[2] another/file.js
<content>

---

3. LLM Compact Mode (--llm-compact)

filemerger src/ --llm-compact

Use this when:

• Token limits are tight
• The project is very large
• Maximum efficiency matters

Characteristics:

• Same structure as --llm
• Fewer blank lines
• Minimal formatting overhead

---

4. Statistics

Use --stats to print merge statistics:

filemerger src/ --stats

Reported values:

• Number of files
• Total lines
• Total bytes
• Skipped files (binary / non-UTF8)

---

5. AI Marker Mode (--ai-markers)

filemerger src/ --ai-markers

Use this when:

• You need strong, explicit file boundaries for AI systems
• You want deterministic multi-file reasoning
• You are feeding large structured context into LLMs
• You need machine-parsable output

Characteristics:

• Explicit file boundary markers
• Clear begin/end delimiters
• Unambiguous separation between files
• Designed for reliable AI ingestion

Example:

<<<FILE 1: path/to/file.py>>>
<content>
<<<END FILE>>>

<<<FILE 2: another/file.js>>>
<content>
<<<END FILE>>>

---

Configuration (Optional)

FileMerger supports an optional .filemerger.toml file in the project root.

Example:

[filters]
max_file_size_mb = 1
exclude_dirs = ["tests"]

[output]
separator_length = 60

If the file is not present, default behavior is used.

---

License

This project is licensed under the MIT License. See the LICENSE file for details.