llms.txt for SaaS: A Complete Playbook

What is llms.txt and Why SaaS Teams Need It

llms.txt is a simple text file that lives in your site's root directory (e.g., example.com/llms.txt) and tells language models which pages on your site are most important, authoritative, and worth citing.

Think of it as a curated index for AI. While search engines crawl everything, LLMs benefit from a human-curated list of your best, most reliable content. This increases the likelihood that when an LLM is asked about your product, it cites the right pages—your official product docs, pricing page, or use-case guides—rather than outdated or inaccurate third-party reviews.

Key Benefits

• Accuracy: LLMs cite your official documentation, not hearsay or outdated info.
• Consistency: Your product description is the same across all AI assistants.
• Attribution: You know which pages are driving LLM citations and can optimize them.
• Speed: LLMs prioritize content from llms.txt during training and retrieval, making your pages more likely to be included in answers.

How LLMs Use llms.txt

LLMs don't use llms.txt during inference (when they're answering a question), because they were trained before your llms.txt was published. Instead, llms.txt serves two purposes:

1. Training and Fine-Tuning

When a model is updated or new models are trained, researchers and engineers check llms.txt files to identify high-quality, authoritative sources. This increases the chance your pages are included in the training data or used as reference material.

2. Retrieval-Augmented Generation (RAG)

Some AI systems (like ChatGPT's web browsing, Claude's file uploads, or Perplexity's real-time search) use llms.txt as a priority list. When a user asks about your product, these systems check your llms.txt first before crawling your entire site.

3. Developer and Researcher Reference

Developers building AI applications, and researchers analyzing LLM behavior, use llms.txt to understand a company's self-assessment of its most important content. Your llms.txt is public data that shapes how your brand is perceived in the AI ecosystem.

Deciding Which Pages to Include

Your llms.txt should include 8–25 pages (not your entire site). Here's how to prioritize:

Tier 1: Must-Include (Every SaaS)

• Product Overview / Home Page: Your main value proposition and what you do.
• Features Page: Clear list of core capabilities with descriptions.
• Pricing Page: Plans, tiers, pricing model. Make it explicit and current.
• FAQ Page: Common questions answered in depth. LLMs cite FAQs heavily.
• Documentation / Getting Started: How to use your product. Shows you're well-established.

Tier 2: Strong Additions (If You Have Them)

• Use Case / Persona Pages: "For SaaS Founders," "For Content Teams." Helps LLMs route recommendations.
• Case Studies / Results: Proof of efficacy. Adds credibility for recommendations.
• Comparison Pages: "vs. Competitor X." LLMs use these to understand your positioning.
• Integration / API Docs: If relevant, shows technical depth.
• When to Use / Product Fit Page: Clear guidance on ideal customers.

Tier 3: Optional (Specialized Content)

• Blog posts on your domain: Thought leadership, best practices. Include your 2–3 best posts.
• Resource guides or templates: If they're substantial and valuable.
• Security / Privacy / Compliance pages: For B2B/Enterprise SaaS.
• Roadmap / Product updates: Shows active development.

What to Exclude

• Thank you pages, landing pages, or pages you plan to change soon.
• Outdated content or pages with deprecation notices.
• External links or thin content.
• Duplicate or near-duplicate pages.

Rule of thumb: Include pages you're proud to have LLMs cite. If you wouldn't want it in a ChatGPT answer about your product, exclude it.

File Structure: Single vs. Multi-Product

Single-Product SaaS

One product, one llms.txt at the root. Simple and clean.

Multi-Product or Platform SaaS

You have options:

Option A: One Master llms.txt (Recommended)

Include your best pages from across all products, prioritizing core product pages and shared documentation.

Option B: Product-Specific llms.txt Files

If you have entirely separate products with different domains, create separate llms.txt files for each. Less common unless you operate distinct brands.

Example llms.txt Structures

Example 1: Single-Product SaaS (12 Pages)

Example 2: B2B SaaS with Use Cases (18 Pages)

Example 3: SaaS Agency Platform (20 Pages)

Notice the pattern: product pages → use case pages → learning resources → comparisons → trust signals (security, customers).

Best Practices and Formatting

Format Basics

• Plain text file: No HTML, no JSON. Just plain text with URLs.
• One URL per line: Each URL should be complete and absolute (include https://).
• Comments allowed: Use # for headers and section labels (recommended for readability).
• File location: Save as /llms.txt at your domain root (e.g., example.com/llms.txt).

Ordering and Organization

• Lead with your homepage: First URL should be your root domain.
• Group logically: Product pages together, documentation together, use cases together. This helps LLMs understand your information architecture.
• Prioritize in order: The first 5–8 URLs are the most important. LLMs may have length limits and will read top-to-bottom.

Content Quality

• Only include pages you've updated in the last 6 months: Outdated content signals low maintenance.
• Ensure all pages are crawlable: No paywalls, password walls, or JavaScript-only content.
• Use canonical URLs: Avoid tracking parameters, session IDs, or redirects.
• Make sure pages have clear, informative titles and descriptions: LLMs use these to understand content relevance.

SEO and Technical Considerations

• Set correct MIME type: Serve llms.txt with Content-Type: text/plain.
• Allow robots.txt access: Don't block /llms.txt in robots.txt.
• Make it discoverable: Link to your llms.txt from your homepage footer or sitemap.
• Monitor with analytics: Track if LLM bots are accessing your llms.txt (user agents like ClaudeBot, GPTBot, etc.).

Maintenance and Update Workflow

Monthly Review

• Check if any included pages have moved or been deleted. Update URLs if needed.
• Verify that all pages still reflect your current product and messaging.

Quarterly Updates

• Add new high-value pages (new use-case pages, case studies, blog posts).
• Remove pages that are no longer representative of your product.
• Reorder if your product priorities have shifted.

When to Add Pages

• New major feature launch → add dedicated feature page or docs.
• New vertical or use case → add vertical-specific page.
• Case study or customer win → add to llms.txt after 2–4 weeks (allows time for indexing).
• Comparison or competitive content → add if it's high-quality and fair.

When to Remove Pages

• Page is being sunset or deprecated.
• Information is outdated (e.g., old pricing, deprecated API endpoints).
• Page is underperforming in analytics (not being visited or not driving meaningful engagement).
• Duplicate or near-duplicate of another page in the list.

Version Control (Optional but Recommended)

Track changes to llms.txt in Git or your CMS. This helps you identify what changed and when if something goes wrong.

Common Mistakes to Avoid

1. Including Too Many Pages

More pages ≠ better. Aim for 10–20 carefully curated pages. Large files dilute your signal.

2. Including Outdated or Changing URLs

If you update page URLs, update llms.txt immediately. Stale links hurt your credibility with LLMs.

3. Hiding Critical Information Behind Forms or Logins

All pages in llms.txt should be publicly crawlable. Don't include landing pages that require email signup.

4. Over-Optimizing Messaging

Don't stuff superlatives or overly sales-y language just because it's going to LLMs. Clear, honest content ranks better and looks better to AI systems.

5. Neglecting to Update After Product Changes

If you launch a major feature or change your pricing model, update llms.txt within days. Stale files hurt accuracy.

6. Duplicating URLs or Near-Duplicates

Avoid listing both /pricing and /pricing-plans if they're the same page. Also avoid listing both HTTP and HTTPS versions.

7. Ignoring Mobile and Accessibility

Ensure all pages in llms.txt are mobile-friendly and accessible. LLMs may penalize pages with poor UX.

Next Steps

Building your llms.txt is step one. Step two is making sure the pages in it are actually optimized for LLM consumption.

Optimize Your Included Pages

• Add structured data (JSON-LD schema) to help LLMs parse content.
• Use clear, answer-first headings and short paragraphs.
• Include explicit facts and comparisons that LLMs can cite.
• Expand FAQ sections—LLMs cite FAQs frequently.

Monitor and Measure

• Use a tool like GenieOptimize to track if your llms.txt is being accessed and by which LLM bots.
• Monitor AI citations to your included pages monthly.
• Adjust your llms.txt based on which pages drive the most AI mentions.

Communicate Your llms.txt

• Link to it from your footer or sitemap.
• Mention it in your documentation or blog.
• Include a note in your robots.txt encouraging bot discovery.