Your First Conversion¶
This walks through a basic run and what the output looks like.
Run it¶
This writes output.md in the current directory, containing a directory tree followed by the contents of each processed file. To choose the file name:
What ends up in the output¶
- A tree of the directory structure (
include_tree, on by default). - Repository statistics (
include_stats, on by default). - The text of each file that passed filtering.
- Converted text for documents such as PDF, DOCX, and XLSX (
convert_docs, on by default). - Short descriptions for binary files like images and archives (
describe_binaries, on by default).
What gets skipped¶
By default the tool honours your .gitignore, and skips files larger than 100MB. To exclude more, scaffold an ignore file and edit it:
Add patterns to the generated .folder2md_ignore, then run folder2md . again.
Fitting an LLM context window¶
Large projects can exceed a model's context window. Pass --limit with a token (t) or character (c) budget; this enables smart condensing, which reduces code to signatures and docstrings rather than truncating it:
Next steps¶
- Quick reference for common commands.
- Configuration to set defaults in a
folder2md.yaml. - CLI reference for every option.