This is interesting. Just wondering about your traffic volume and how long you have been running lcoalpdf?
For us, it is more like 5% of the traffic from GEO, but we have been running the company for 2 years and have created a lot of handwritten content for devs.
Volume is modest (~180 visitors/month), but the 50/50 split is what's interesting.
Been in production since August 2025, so ~4 months.
The strategy was intentional from the start: there's no point competing with Adobe, Smallpdf, ILovePDF for Google rankings. They have 10+ years of backlinks, massive marketing budgets, and domain authority I'll never match as a solo dev.
So I made a bet on GEO from day one:
- Semantic HTML that LLMs can parse
- Clear technical docs (GitHub README as primary content)
- Honest about limitations
- Privacy-first architecture (client-side processing)
Your 5% GEO makes sense for a 2-year-old company optimizing for traditional SEO. The difference: I skipped the SEO game entirely. When you're competing in an established niche, GEO-first might be the only viable strategy for bootstrapped products.
Curious: what type of dev content are you creating? And have you tested how LLMs cite it vs your traditional marketing content?
Hi HN! I'm the creator of LocalPDF (https://localpdf.online) — a privacy-first PDF toolkit that works entirely in your browser.
I wanted to share something fascinating: ChatGPT has become my biggest traffic source, accounting for ~50% of visitors, compared to ~45% from Google. This happened organically — I didn't optimize specifically for AI assistants.
Key Stats:
- ChatGPT: ~50% of traffic
- Google Search: ~45%
- Direct/Other: ~5%
- Conversion rate from ChatGPT: 2x higher than Google
- ChatGPT users stay 40% longer on average
What I think is happening:
GEO (Generative Engine Optimization) vs SEO
Traditional SEO focuses on keywords, backlinks, and content clusters. But AI assistants work differently:
- They parse semantic HTML structure
- They value clear, descriptive content over keyword density
- They prefer sites that solve problems directly
The ironic part? My site was built with basic semantic HTML and clear descriptions — not because I was targeting AI, but because it's good web development practice.
Technical Details:
- Stack: Astro + TypeScript, client-side PDF.js
- No tracking, no uploads — everything processes locally
- Semantic HTML: proper headings, landmarks, meta descriptions
- Clear tool descriptions with use cases
Why ChatGPT users convert better:
1. Intent clarity: They ask specific questions ("how to compress PDF without losing quality")
2. Contextual recommendations: ChatGPT explains why LocalPDF fits their needs
3. Trust transfer: AI assistant endorsement acts as social proof
Limitations:
- Can't track which specific ChatGPT prompts drive traffic
- No direct way to optimize for AI recommendations
- ChatGPT's recommendations can change unpredictably
Questions for HN:
1. Are you seeing similar traffic patterns from AI assistants?
2. How do you measure/optimize for GEO vs traditional SEO?
3. What happens when every AI assistant has different ranking criteria?
This feels like early Google days — there's a level playing field where good products with clear value propositions can get discovered organically, without massive SEO budgets.
Happy to answer questions about the technical implementation, traffic patterns, or anything else!
Petition to start using AIO instead of GEO and not butcher geospatial/geographic/maps industries like we did with crypto for crypto currency and cryptography _/\_
This is interesting. Just wondering about your traffic volume and how long you have been running lcoalpdf?
For us, it is more like 5% of the traffic from GEO, but we have been running the company for 2 years and have created a lot of handwritten content for devs.
Volume is modest (~180 visitors/month), but the 50/50 split is what's interesting.
Been in production since August 2025, so ~4 months.
The strategy was intentional from the start: there's no point competing with Adobe, Smallpdf, ILovePDF for Google rankings. They have 10+ years of backlinks, massive marketing budgets, and domain authority I'll never match as a solo dev.
So I made a bet on GEO from day one: - Semantic HTML that LLMs can parse - Clear technical docs (GitHub README as primary content) - Honest about limitations - Privacy-first architecture (client-side processing)
Your 5% GEO makes sense for a 2-year-old company optimizing for traditional SEO. The difference: I skipped the SEO game entirely. When you're competing in an established niche, GEO-first might be the only viable strategy for bootstrapped products.
Curious: what type of dev content are you creating? And have you tested how LLMs cite it vs your traditional marketing content?
Hi HN! I'm the creator of LocalPDF (https://localpdf.online) — a privacy-first PDF toolkit that works entirely in your browser.
I wanted to share something fascinating: ChatGPT has become my biggest traffic source, accounting for ~50% of visitors, compared to ~45% from Google. This happened organically — I didn't optimize specifically for AI assistants.
Key Stats: - ChatGPT: ~50% of traffic - Google Search: ~45% - Direct/Other: ~5% - Conversion rate from ChatGPT: 2x higher than Google - ChatGPT users stay 40% longer on average
What I think is happening:
GEO (Generative Engine Optimization) vs SEO Traditional SEO focuses on keywords, backlinks, and content clusters. But AI assistants work differently: - They parse semantic HTML structure - They value clear, descriptive content over keyword density - They prefer sites that solve problems directly
The ironic part? My site was built with basic semantic HTML and clear descriptions — not because I was targeting AI, but because it's good web development practice.
Technical Details: - Stack: Astro + TypeScript, client-side PDF.js - No tracking, no uploads — everything processes locally - Semantic HTML: proper headings, landmarks, meta descriptions - Clear tool descriptions with use cases
Why ChatGPT users convert better: 1. Intent clarity: They ask specific questions ("how to compress PDF without losing quality") 2. Contextual recommendations: ChatGPT explains why LocalPDF fits their needs 3. Trust transfer: AI assistant endorsement acts as social proof
Limitations: - Can't track which specific ChatGPT prompts drive traffic - No direct way to optimize for AI recommendations - ChatGPT's recommendations can change unpredictably
Questions for HN: 1. Are you seeing similar traffic patterns from AI assistants? 2. How do you measure/optimize for GEO vs traditional SEO? 3. What happens when every AI assistant has different ranking criteria?
This feels like early Google days — there's a level playing field where good products with clear value propositions can get discovered organically, without massive SEO budgets.
Happy to answer questions about the technical implementation, traffic patterns, or anything else!
Petition to start using AIO instead of GEO and not butcher geospatial/geographic/maps industries like we did with crypto for crypto currency and cryptography _/\_
Ha! AIO makes sense. Though knowing tech, we'll probably end up with 5 competing acronyms
I thought people were already using AEO (answer engine optimization)...
Note to self. Pick a product name that an LLM could hallucinate.
Adversarial branding is going to be a real niche in 2026.