Suggestions
LLMs.txt Lab logo LLMs.txt Lab
Examples Tutorials WordPress Plugin
All Tools
llms.txt Generator & Builder llms.txt Validator llms.txt Formatter llms.txt Converter llms.txt Link Checker llms.txt Analyzer & Previewer llms.txt Merge & Compare Tool llms.txt Split Tool

llms.txt Split Tool

Split a large llms.txt file into smaller, organized files by section, category, size, link count or estimated tokens while preserving its Markdown structure.

1
Input File
2
Configure Split
3
Review & Export

Paste Content

Click to upload or drag file

Supports .txt and .md

What Is the llms.txt Split Tool?

The llms.txt Split Tool is designed to divide large or multi-topic llms.txt files into smaller, organized components without breaking their internal structure. Because large documentation hubs often exceed standard context limits, splitting files intelligently—by specific sections, link count, or total size—is crucial. This tool parses your Markdown and carefully separates blocks while preserving titles, summaries, valid URLs, and optional classifications.

How the Split Tool Works

Unlike raw text splitters that indiscriminately cut content based on characters, this tool constructs an internal model of your llms.txt file before splitting. This ensures that a Markdown link and its accompanying description are never sliced in half. You can configure the engine to output independent downloadable chunks for local tooling or scoped website files (like /docs/llms.txt) along with a root index file for immediate deployment.

When Should You Split an llms.txt File?

You should consider splitting your file when your documentation spans multiple vast subjects (e.g., separating user tutorials from dense API references) or when the single file consistently exceeds the context windows of average LLMs. It is also beneficial when migrating to a complex, scoped architecture like Cloudflare’s model, where a main root file links to highly specific sub-files.

How to Split by H2 Section

Splitting by H2 Section is the safest and most natural method. The tool automatically detects every ## Section Name and places its contents into an independent output file. It intelligently prepends the project title to ensure each new file remains compliant with the single H1 requirement of the specification.

How to Split by Link Count

If a specific section is overwhelmingly large, you can split by a maximum link threshold. The engine will chunk the section sequentially into parts, keeping all descriptions tied to their parent URLs and ensuring the H2 section title persists at the top of each new file block.

How to Split by File Size

Using the File Size method allows you to enforce strict byte limits per file. The tool calculates UTF-8 sizes and divides the document at the closest safe Markdown boundary, ensuring it never truncates fenced code blocks, summaries, or internal file lists mid-sentence.

How Token-Based Splitting Works

Token estimation uses a rough heuristic (typically ~4 characters per token) to help you gauge the context weight of the output. While different tokenizers behave differently, this provides an excellent baseline for ensuring your individual llms.txt chunks do not blindly overflow a target model’s context limit.

How Scoped llms.txt Files Work

Instead of generic chunk names (e.g., part-01.txt), selecting "Scoped Website Files" prepares your files for live deployment. It names the files based on their content (e.g., api-reference.txt) which you can map to /api/llms.txt on your server, ensuring LLMs can traverse your documentation structurally.

How the Root Index Is Generated

If you output scoped files, the tool can generate a root llms.txt file. This index file will contain standard Markdown links pointing to your child files. This is exactly how massive ecosystems manage discovering nested, topic-specific context files safely.

How the Optional Section Is Handled

The ## Optional section holds secondary priority in the official spec. When splitting, the tool can actively recognize this specific heading and ensure it remains separated into an independent chunk, ensuring that lightweight LLM crawlers can safely ignore it when building tight context windows.

What Information Is Preserved?

The engine preserves your single H1 project title, any blockquote summaries, introductory paragraphs, all H2 boundaries, every Markdown link, trailing descriptions, and URL structure. When splitting large sections, it duplicates structural wrappers (like the H2 name) into child chunks so context is never orphaned.

How to Deploy Split llms.txt Files

Once you download the generated ZIP, extract the files to your local repository. Upload the root llms.txt to your public root (/public_html/llms.txt) and place the scoped child files into their respective subdirectories (/public_html/api/llms.txt). AI crawlers will then parse the root file to discover the children.

Split Tool vs Merge Tool

The Split Tool divides massive unified files into a manageable tree of structured child elements. Conversely, the llms.txt Merge & Compare Tool resolves differences between two files, deduplicates URLs, handles conflict resolution, and compiles them safely back into a single unified output.

Related Components

llms.txt Generator & Builder Create an llms.txt file manually, from a sitemap or from a list of URLs. llms.txt Validator Check syntax, Markdown and structure. llms.txt Formatter Clean, beautify and reorder sections. llms.txt Converter Convert between CSV, JSON and HTML. llms.txt Link Checker Find broken and duplicate URLs. llms.txt Analyzer & Previewer Preview, URLs, sections and word count.

Questions you might have

Does the llms.txt specification require large files to be split?
No. The proposal permits files at the root or in subpaths but does not define a mandatory file size or splitting threshold.
Can I create a separate llms.txt for each website section?
Yes. The proposal permits files in subpaths. A main root file should link clearly to scoped files because automatic subpath discovery is not fully standardized.
Will the tool change my links?
No. The default splitting process should preserve URLs, titles and descriptions exactly.
Is the project H1 copied into every file?
Every independently valid child file should have an H1. The tool can use the original title or combine it with the child scope.
What happens to the summary?
The user can copy it, keep it only in the root, or create scoped summaries. The summary is optional under the current format.
What happens to the Optional section?
It can remain in a child file, be split into its own file, or be represented through an Optional link in the root index. The original Optional classification should not be lost.
Are token limits official?
No. Token targets are processing preferences. Counts also vary by model and tokenizer.
Can the tool split one large section?
Yes. It can divide that section by link count, size, characters or estimated tokens while preserving complete link records.
Does the tool download all files together?
Yes. It should provide a ZIP containing child files, the optional root index, a manifest and deployment instructions.
Can I merge the files again?
The split-project manifest can preserve enough information to reconstruct the original ordering. Your Merge & Compare Tool can also combine the generated files.
llms.txt Merge & Compare Tool Merge tool and diff checker. llms.txt Generator & Builder Create an llms.txt file manually, from a sitemap or from a list of URLs.