Photo by Rahul Mishra on Unsplash
73% of brands have zero mentions in AI-generated responses — even when they rank on page one of Google. For developers, that gap is getting expensive. Perplexity, ChatGPT search, AI coding agents, and AI recruiting tools are increasingly the first contact when someone wants to know who you are or what you've built. If your profile doesn't speak their language, you don't show up.
The fix is a 25-line markdown file called llms.txt. Setting up your llms.txt developer profile takes 30 minutes and puts you in front of AI tools that most developers aren't visible to yet.
What llms.txt Is — and Why Your GitHub README Doesn't Cut It#
llms.txt is a plain-text markdown file at **yoursite.com/llms.txt** that tells AI tools, in structured form, who you are and what content they'll find. Think of it as robots.txt for AI — but instead of telling crawlers what not to access, it tells them what's worth reading.
The standard was proposed in September 2024 by Jeremy Howard of Answer.AI, who described the core problem plainly: "context windows are too small to handle most websites in their entirety," and converting HTML pages cluttered with navigation, ads, and JavaScript into something an LLM can parse is "both difficult and imprecise." llms.txt solves that by giving AI tools a single, structured entry point.
By July 2025, 600+ sites had adopted the standard — Anthropic, Stripe, Vercel, Cursor, Hugging Face, Cloudflare, and Zapier among them. Chrome Lighthouse 13.3 (released May 2026) now audits for the file.
Your GitHub README doesn't do any of this. It's a human document, hosted on GitHub's domain, written for visitors who navigate there intentionally. When Perplexity researches your name or ChatGPT answers "who are the best indie hackers building with Next.js," it doesn't browse to your GitHub profile. It queries the web, finds your personal site or profile, and processes whatever it can parse. A README can't be found that way. A well-formed llms.txt can.
The llms.txt spec explicitly lists "personal CVs" as one of its core use cases — alongside product documentation, e-commerce catalogs, and legislative resources. It was built for exactly this.
The AI Hiring Shift: Why This Matters Now#
84% of developers use or plan to use AI tools in their workflow (Stack Overflow Developer Survey 2026). The flip side: 87% of companies now use AI-driven hiring tools, and those tools reach 89–94% accuracy on resume parsing and skill matching — but only when they have structured data to work with.
The same dynamic applies to the unofficial discovery layer. When a recruiter, potential cofounder, or client asks Perplexity or ChatGPT to find developers with a specific stack, the AI synthesizes from whatever it can read on the open web. Over 73% of brands are invisible in AI-generated responses even while ranking on page one of Google. The correlation between traditional SEO visibility and AI visibility doesn't hold. Structured, machine-readable content wins.
This matters for individual developers, not just companies. 36% of developers reported learning AI-enabled tools specifically to advance their career in the last year. llms.txt is the next step: not just using AI, but being discoverable by it.
And the timing is right to move early. Only 10.13% of domains have adopted llms.txt so far. Among developer-facing SaaS companies (Stripe, Cloudflare, Cursor), adoption is now routine. Among individual developers, it's essentially zero. First-mover advantage on an emerging standard is real — and the cost of shipping your first llms.txt is measured in minutes.
The AI-Readable Profile Stack#
A developer profile that works in 2026 has three distinct layers, each targeting a different reader:
Layer 1 — Human-readable: Your bio page at devbio.me/yourname. This is where people land. It shows your GitHub activity, live MRR, project cards with real traction numbers, skills, and links — all in a composable layout built for scanning.
Layer 2 — Machine-readable: Your /llms.txt. This is what AI tools process. It summarizes your identity in structured markdown, links to your key content, and speaks directly to LLMs, coding agents, and AI search at inference time.
Layer 3 — ATS-readable: Your /yourname/resume PDF — compiled from your live profile data, formatted in LaTeX for applicant tracking systems. This is what hiring software ingests during the formal application process.
Most developers have Layer 1 covered, sometimes Layer 3. Layer 2 is the gap — and it's the layer that feeds discovery before anyone visits your page at all.
The three layers aren't redundant. Your bio converts people who find you. Your resume gets you through hiring systems. Your llms.txt is how the right AI tools find you in the first place. Each does a distinct job for a distinct reader.
For a deeper look at what goes on Layer 1, the developer bio components guide breaks down every section that builds credibility — from GitHub stats to live revenue to the about section that actually gets read.
What Your llms.txt Developer Profile Should Include#
The spec requires only one element: an H1 heading with your name or project. Everything else is structured and modular. For a developer profile, a well-formed llms.txt looks like this:
# Jane Kim — Full-Stack Developer & Indie Hacker
> Independent developer specializing in Next.js, TypeScript, and PostgreSQL.
> Building SaaS products since 2022. Current MRR: $4,200. Based in Seoul.
> GitHub: 847 contributions in the last year, 12 public repos.
## Skills
- Next.js, TypeScript, React, PostgreSQL, Redis, Tailwind CSS, Node.js
## Projects
- [Formblast](https://formblast.io): Form backend SaaS. 340 paying customers, $4,200 MRR.
- [Pagecrawl](https://pagecrawl.dev): Automated web change monitoring. Open source, 1.2k GitHub stars.
## Experience
- Senior Engineer, Naver (2019–2022) — Search infrastructure team
- Independent developer (2022–present)
## Links
- Bio: https://devbio.me/janekim
- GitHub: https://github.com/janekim
- Resume: https://devbio.me/janekim/resume
- LinkedIn: https://linkedin.com/in/janekimThe format follows the spec: H1 identity, a blockquote summary that AI tools parse first, then H2 sections for skills, projects, experience, and links. This is the complete llms.txt developer profile structure.
Two things make this specifically powerful for developers:
The blockquote is a citation target. When ChatGPT or Perplexity answers "indie developers building SaaS with TypeScript in Korea," your summary blockquote is the text it lifts. Write it factual and specific — "independent developer specializing in Next.js, TypeScript, and PostgreSQL. Current MRR: $4,200" beats "passionate developer who loves building things." Vague claims don't get cited. Specific facts do.
Revenue is a rare proof signal. "Current MRR: $4,200" in your summary is verifiable, specific, and unusual. Most developer profiles claim traction. Very few quantify it in a machine-readable format that AI tools can read at inference time. If you're building in public and sharing revenue numbers anyway, this is the highest-leverage place to put them.
If each project has its own llms.txt at its domain root, link to those too — the spec allows chaining. An AI reading your profile llms.txt can pull deeper project context automatically from each project's file.
How AI Tools Use Your llms.txt Right Now#
The honest answer: retrieval at inference time, not training. No major LLM has publicly committed to reading llms.txt in their training pipeline. What's confirmed and measurable is what happens during a live query.
When Perplexity, ChatGPT search, or an AI agent queries the web, it fetches pages and processes their content. A well-formed llms.txt at your domain root is findable, clean for parsing, and optimized for exactly the context-window constraint that makes full HTML pages hard to use. The same mechanism runs in AI coding agents like Cursor and GitHub Copilot — they fetch llms.txt from libraries at inference time, right now, to understand APIs they're helping developers build with.
One documented case study showed submitting an llms.txt to Google Search Console led to AI-powered search answers within 72 hours. That's the clearest evidence of measurable discoverability impact.
The adoption curve matters here. Perplexity, OpenAI, Anthropic, Stripe, Cloudflare, Microsoft Teams SDK — all use llms.txt as a routing layer for AI coding agents. These aren't experiments. They're production infrastructure for developer-facing AI. Individual developer profiles that adopt the same format put themselves in the same discovery pool.
llms.txt vs. robots.txt vs. sitemap.xml vs. README#
These files aren't competitors — they're layers for different readers. Most developer profiles are missing the AI layer entirely.
File | Who reads it | Purpose | Developer profile use |
|---|---|---|---|
| Search crawlers | Control what bots access | Not applicable |
| Search crawlers | Index your URLs | Not applicable |
| AI tools, LLMs, coding agents | Summarize identity + key content | Primary AI discovery |
GitHub | Humans on GitHub | Narrative profile | Human-only, GitHub-scoped |
| ATS systems, hiring managers | Structured work history | Hiring-specific, not web-discoverable |
A complete developer profile serves all of these readers. robots.txt and sitemap.xml are table stakes for any site. Your GitHub README is the human-facing layer on GitHub's domain. Your llms.txt is the AI discovery layer on your domain. Your resume PDF handles ATS. Each targets a different system, at a different point in the discovery-to-hire funnel.
The mistake most developers make is treating their GitHub profile as their entire online presence. It isn't — it's one layer, for one reader type, on one domain. The case for going beyond the README covers this in detail: 82% of your GitHub work is in private repos that don't show on your profile at all.
Who's Already Shipping llms.txt#
The adoption list is its own argument. By July 2025, over 600 sites had implemented llms.txt. The early movers:
Anthropic and Stripe — core AI developer infrastructure. They ship llms.txt because their users (developers) use AI coding agents to build with their APIs, and those agents fetch llms.txt at inference time.
Cursor and GitHub Copilot — AI coding environments that pull llms.txt from libraries and tools in real time during development sessions.
Hugging Face, ElevenLabs, Zapier — AI-native products that understand the discovery layer earlier than most.
Cloudflare — infrastructure at massive scale; their llms.txt adoption signals that it's worth maintaining for long-term AI indexing.
Yoast SEO — when a WordPress SEO plugin adds a feature, that feature has crossed into mainstream publishing tooling.
The pattern is consistent: developer-tool companies adopted first, because their audience — developers — uses AI to research both tools and people. Individual developers are the next wave. If Anthropic thinks it's worth maintaining a machine-readable summary of who they are and what they do, it's worth 30 minutes of your time to do the same.
How DevBio Generates Your llms.txt Automatically#
Building an llms.txt from scratch is straightforward — the template above shows it's not complex. The harder problem is keeping it current. Every project you ship, every MRR update, every stack change makes your llms.txt stale the moment you forget to sync it.
DevBio generates a spec-compliant llms.txt from your profile data automatically. It's available at devbio.me/yourname/llms.txt — or yourname.com/llms.txt if you're on a custom domain. It pulls from the same source of truth as your live profile:
Skills — from your skills component
Projects — each project with live GitHub stars, commit count, and MRR if a payment provider is connected
Experience — from your work history component
Links — your full configured link set, including GitHub, resume, and social profiles
Revenue — live MRR from your connected payment provider (Stripe, Dodo Payments, Lemon Squeezy, or Polar)
The format follows the llmstxt.org spec, so it's compatible with any AI tool that reads the standard. When you edit your bio, your llms.txt updates. There's no separate file to publish or remember to sync.
This is what makes the three-layer stack coherent. You maintain one profile. It renders as a human-readable bio, a machine-readable llms.txt, and an ATS-ready PDF resume — all from the same underlying data.
Before and After: Same Developer, Two Discovery Outcomes#
Here's a concrete scenario showing why the llms.txt developer profile gap matters.
Without llms.txt:
Marcus is a developer with a solid GitHub presence and a personal site he built in 2024. His site has a bio, a projects section, and links. A startup CTO asks Perplexity: "find independent developers who've shipped SaaS with PostgreSQL and Next.js and have paying customers." Perplexity queries the web, finds Marcus's site, tries to parse the HTML, gets his navigation, a few fragments of text, and a jumbled hero section. It either excludes Marcus or summarizes him with wrong or missing details. He doesn't get the email.
With llms.txt:
Same Marcus, same profile, same CTO query. Perplexity finds his llms.txt. The blockquote reads: "Independent developer. 2 shipped SaaS products. $1,800 MRR. Next.js, PostgreSQL, TypeScript. Open to contract work." Perplexity surfaces Marcus as a specific, credible match — stack confirmed, revenue confirmed, availability stated. The CTO reaches out.
The actual profile content didn't change. The discoverability did.
This scenario plays out for freelance leads, cofounder searches, conference speaker invitations, and open-source collaboration requests — any context where someone uses an AI tool to find a developer with a specific set of criteria. A well-structured llms.txt is what puts you in the results.
Building Your Developer Personal Brand for AI Search#
The broader context here is that developer personal branding is shifting from SEO-optimized pages to entity-optimized profiles. The six factors that determine whether AI systems cite you are: retrieval-ready content structure, entity signal strength, third-party citation presence, platform-specific optimization, technical infrastructure performance, and measurement infrastructure.
llms.txt directly addresses the first two. A profile that has a clean llms.txt, a strong structured bio, and external links (GitHub, published work, talks) covers the entity signal layer that makes AI tools confident enough to surface you in a response.
For developers, this is a compounding advantage. Your developer personal brand isn't just about how humans perceive you — it's about how AI systems classify and recall you. A developer who's been building in public, shipping products with real revenue, and maintaining a machine-readable profile is the exact kind of signal that AI tools flag as credible.
The 4-layer developer presence framework covers this in full — human-readable, machine-readable, ATS-readable, and publicly verifiable. llms.txt is the machine-readable layer that most developers are missing.
Frequently Asked Questions About the llms.txt Developer Profile#
What is llms.txt and who reads it?
llms.txt is a plain-text markdown file at the root of your website, proposed by Jeremy Howard of Answer.AI in September 2024. It gives LLMs, AI search tools like Perplexity, and AI coding agents like Cursor a structured, scannable summary of who you are and what content they'll find on your site. It solves the context-window problem: AI tools can't process full websites in their entirety, but they can process a well-formed 30-line markdown file in seconds. For developers, the llms.txt developer profile use case is explicitly included in the original spec.
Does Google index llms.txt?
Google's standard web crawlers index it as a regular text file — it shows up in Search Console alongside your other pages. But llms.txt isn't a substitute for traditional SEO. Chrome Lighthouse 13.3 (May 2026) added an audit check for the file, signaling that Google's toolchain treats it as a meaningful signal for AI-optimized content. Its primary value today is AI tool discoverability — Perplexity, ChatGPT search, and AI agents at inference time — not traditional search ranking.
What should a developer's llms.txt include?
At minimum: an H1 with your name and role, a blockquote summary with your stack, MRR or notable projects, and a factual description of what you do, then H2 sections for Skills, Projects, Experience, and Links. Revenue and GitHub stats in the blockquote make a real difference — AI tools surface specific, verifiable claims far more often than vague ones. "Next.js, TypeScript, $2,100 MRR" is far more citable than "passionate full-stack developer."
Is the llms.txt standard official?
It's a proposed standard from llmstxt.org, not an IETF or W3C ratified spec. Adoption is organic. By July 2025, 600+ websites had implemented it, including Anthropic, Stripe, Vercel, Cursor, and Cloudflare. Chrome Lighthouse 13.3 added auditing for the file in May 2026. No major LLM provider has committed to using it in training pipelines, but retrieval at inference time (AI search, AI agents) is documented and measurable today.
How is llms.txt different from a sitemap or robots.txt?
robots.txt tells crawlers what not to access. A sitemap lists all your URLs for search indexing. llms.txt tells AI tools who you are and what content is worth reading — it's content-level and identity-level, not URL-level. All three can coexist and serve different readers. A developer profile needs all of them, but llms.txt is the only file specifically designed for AI tool discovery at inference time.
How often should I update my llms.txt?
Every time you ship a project, change your primary stack, or your revenue numbers shift significantly. Stale data is worse than no file — AI tools may surface outdated or contradictory information. The strongest argument for auto-generated llms.txt (like DevBio's) is that it stays in sync with your live profile automatically. You edit your bio once, all three formats update.
Can I link to my project's llms.txt from my profile llms.txt?
Yes — the spec explicitly supports this. Link to each project's llms.txt URL in your projects section: [ProjectName](https://project.com/llms.txt): brief description. This creates a chain of discovery: an AI reading your profile llms.txt can fetch richer context about each project automatically. It's the same pattern Stripe uses — their main llms.txt links to individual product llms files.
What's the fastest way to get an llms.txt for my developer profile?
If you're on DevBio, visit devbio.me/yourname/llms.txt — it's already generated from your profile data and updates automatically. If you're rolling your own, the template in this post is the full spec: H1 name and role, blockquote summary with stack and revenue, then Skills, Projects, Experience, and Links sections. Keep the blockquote under 3 sentences and make every sentence factual and specific.
Your Profile, Machine-Readable#
The developer discovery layer is shifting. Traditional search ranks pages. AI tools synthesize profiles. If yours is HTML-only, it's invisible to an increasingly large share of the first-contact layer — the AI tools that people use before they ever visit a site.
Three things to do this week:
Check
**devbio.me/yourname/llms.txt**— if you're on DevBio, it's already there. Review your blockquote summary for specificity.Strengthen your blockquote — include your primary stack, one or two shipped products, and a revenue number if you have one. Specific facts get cited. Vague claims don't.
Link to your llms.txt from your bio's links section, so AI tools can find it in one hop from any page that references you.
The adoption window is still early. Only 10% of domains have it. Essentially no individual developers do. Being machine-readable before your peers are is exactly the kind of compounding advantage that's worth acting on now.
Your code already proves you can build. Put it in a format AI can read — devbio.me.