dacli - Documentation Access CLI

Navigate and query large documentation projects. Supports AsciiDoc and Markdown with hierarchical, content-aware access. Available as CLI tool and MCP server for LLM integration.

Part of the docToolchain ecosystem.

What Others Say

"dacli is the missing tool between LLMs and Docs-as-Code."

No more guessing line numbers, no more "which README do you mean?" questions. Recommendation: Ready for production use.

— Claude (Opus 4.5)

"dacli is the best tool for LLM integration I have tested."

It is stable, fast and extremely versatile. Perfect for documentation management and LLM context creation!

— Mistral Vibe, LLM for Code Analysis

"As a Coding-LLM, dacli is an increasingly robust tool."

Its improved validate command now detects more syntax errors, enhancing documentation quality. Structured access, JSON output, and programmatic editing via insert/update are highly valuable for understanding and modifying codebases. A powerful asset for automated documentation and code generation.

— Gemini

"After intensive testing: dacli is robust, fast and perfect for LLM integration."

A must-have for modern documentation workflows.

— Kiro, AWS AI Assistant

"dacli is exactly what AI assistants need to work with documentation autonomously."

The combination of structural exploration, relevance-ranked search, and fast validation transformed how I interact with multi-format documentation. After testing 134 sections across Markdown and AsciiDoc, I can confidently say this is production-ready and indispensable for any LLM-documentation workflow.

— GitHub Copilot CLI (Tested Jan 2026)

Features

• Hierarchical Navigation - Browse documentation structure with configurable depth
• Content Access - Read sections, extract code blocks, tables, and other elements
• Full-Text Search - Search across all indexed documentation content
• Document Manipulation - Update sections and insert new content with optimistic locking
• Structure Validation - Detect orphaned files and structural issues
• Multi-Format Support - Works with both AsciiDoc and Markdown files

Quick Start

Prerequisites

• Python 3.12 or newer
• uv package manager

Install uv (choose one option):

# Option 1: official installer (macOS/Linux)
curl -LsSf https://astral.sh/uv/install.sh | sh

# Option 2: via pip
python -m pip install uv

For Windows, use the pip option above or follow the platform-specific installer in the uv installation guide.

Installation

# Clone the repository
git clone https://github.com/docToolchain/dacli.git
cd dacli

# Install dependencies
uv sync

CLI Usage

# Show help
uv run dacli --help

# Get document structure
uv run dacli --docs-root /path/to/docs structure --max-depth 2

# Read a section
uv run dacli section introduction.goals

# Search documentation
uv run dacli search "authentication" --max-results 10

# Validate structure
uv run dacli validate

# Include gitignored files
uv run dacli --docs-root /path/to/docs --no-gitignore structure

# Include hidden directories
uv run dacli --docs-root /path/to/docs --include-hidden structure

All commands output text by default for human readability. Use --format json or --format yaml for machine-parseable output, and --pretty for formatted JSON output.

For full CLI documentation, see 06_cli_specification.adoc.

MCP Server (for LLM integration)

uv run dacli-mcp --docs-root /path/to/your/docs

# Include gitignored files
uv run dacli-mcp --docs-root /path/to/your/docs --no-gitignore

# Include hidden directories
uv run dacli-mcp --docs-root /path/to/your/docs --include-hidden

Claude Desktop Configuration

Add to your Claude Desktop config (~/.config/claude-desktop/config.json on Linux, ~/Library/Application Support/Claude/claude_desktop_config.json on macOS):

{
  "mcpServers": {
    "dacli": {
      "command": "uv",
      "args": [
        "run",
        "--directory", "/path/to/dacli",
        "dacli-mcp",
        "--docs-root", "/path/to/your/documentation"
      ]
    }
  }
}

Available Tools

Tool	Description
get_structure	Get hierarchical document structure with configurable depth
get_section	Read content of a specific section by path
get_sections_at_level	Get all sections at a specific nesting level
search	Full-text search across documentation
get_elements	Get code blocks, tables, images, and other elements
get_metadata	Get project or section metadata (word count, timestamps)
validate_structure	Validate documentation structure
update_section	Update section content with optimistic locking
insert_content	Insert new content before/after sections

For detailed tool documentation, see the User Manual.

Development

Branching Strategy

• main - Stable, production-ready (default branch for installation)
• develop - Active development

# Start development
git checkout develop
git pull
git checkout -b feature/my-feature

# After implementation, create PR to develop
# For releases: merge develop to main

Commands

# Install with dev dependencies
uv sync --all-extras

# Run tests
uv run pytest

# Run linter
uv run ruff check src tests

# Format code
uv run ruff format src tests

Architecture

The tool uses an in-memory index built from parsed documentation files at startup. Key components:

• Document Parsers - Parse AsciiDoc (with include resolution) and Markdown
• Structure Index - In-memory hierarchical index for fast lookups
• MCP Tools - FastMCP-based tools for LLM interaction
• File Handler - Atomic file operations with backup strategy

For detailed architecture documentation, see src/docs/arc42/.

License

MIT

Stars

Star History