Use with AI / LLMs

LogLayer for Go ships two reference files designed to be pasted into a chat with Claude, ChatGPT, Cursor, Copilot Chat, or any other LLM-backed coding assistant. They follow the llmstxt.org convention.

File	Purpose	Size
/llms.txt	Concise index: every page on the docs site with a one-line description and link. Use this when you want the model to know the surface area without burning context on full prose.	small
/llms-full.txt	Comprehensive single-file reference: API surface, types, transports, integrations, thread-safety contract, and key patterns inlined. Use this when you want the model to answer detailed questions without browsing.	medium

Both files are kept in sync with the docs site on every release.

How to use them

With Claude / ChatGPT (web)

Paste the contents of llms-full.txt (or its URL) at the start of your conversation:

Use the loglayer-go reference at https://go.loglayer.dev/llms-full.txt as authoritative. Then write me an HTTP handler that derives a per-request logger using integrations/loghttp and logs the request ID, method, and path.

With Cursor / Continue / similar IDE assistants

Add the URL as a context source. Most IDE assistants accept a URL or local file as a "doc source": drop in https://go.loglayer.dev/llms-full.txt and the model will reference it when answering loglayer-related questions.

With Claude Code / Aider / SDK-based tools

Fetch the file once and include it as a system message:

curl -sSL https://go.loglayer.dev/llms-full.txt > .ai/loglayer.txt

Then add .ai/loglayer.txt to your tool's context-files list (or wire it into a custom system prompt).

Sample prompts

Once the reference is loaded, these prompts produce useful output:

• "Show me how to set up loglayer with the structured transport in production and the pretty transport in dev, switched by an env var."
• "Write a Datadog setup that batches every 2 seconds and pipes send errors to a metrics counter."
• "What's the right pattern for a per-request logger in an HTTP handler?"
• "Explain when to use WithFields vs WithMetadata vs WithContext."
• "Write an integration test using transports/testing that verifies my handler logs the request ID."

Why two files?

llms.txt is for cases where context is precious: it's a sitemap that lets the model fetch only what it needs. llms-full.txt is for cases where context is cheap (long-context models, agentic tools): everything is inlined so the model doesn't need to browse.

If you're not sure, start with llms-full.txt. It's the lower-friction option.

Other ways to feed loglayer to an LLM

• Source code at github.com/loglayer/loglayer-go is small enough to fit in most coding assistants' context.
• pkg.go.dev (pkg.go.dev/go.loglayer.dev) renders all GoDoc, including type signatures and doc comments. Useful for fact-checking the model's output.
• This docs site is itself indexed by most search-augmented assistants. Asking "from the loglayer.dev Go docs, ..." often works without any setup.