Piano's product documentation is exported in formats designed for large language models, following the http://llmstxt.org standard. Use these files to ground coding agents, chatbots, and RAG pipelines on accurate, up-to-date information about Analytics, Subscriptions, Audience, and Amplifier.
Looking for API reference docs instead? This site (http://docs.piano.io ) covers product documentation, how-tos, and support FAQs. For SDK/API reference material aimed at developers integrating Piano Analytics, see http://developers.piano.io/analytics/for-ai-agents/ , which offers additional scoped indexes (data collection, Reporting API, access management).
Which file should I use?
Pick the lightest file that gives your agent enough context. The index is cheap; the full corpus is powerful but heavy, and it's large enough that it's split across multiple pages.
-
Explore & route. Let your agent scan the index and fetch only the pages it needs →
llms.txt -
Ingest everything. Load the full corpus in one request when context window allows →
llms-full.txt -
Deep-dive one topic. Fetch a single page's markdown URL from any index entry →
*.md
Available files
llms.txt
Master index of the entire documentation site: every page, across every product (Analytics, Subscriptions, Audience, Amplifier) and languages (English, French, Japanese). Each entry links to a machine-readable .md page.
Best for: starting point for any AI agent, discover topics and route to the right page.
Address: https://docs.piano.io/llms.txt
llms-full.txt
Complete documentation content in a single export, following the http://llmstxt.org standard. Because the corpus spans four products and multiple languages, this file is paginated. Follow the "Next Page" link at the bottom of each page to continue.
Best for: one-shot ingestion when you need broad context and have a large context window.
Addresses:
https://docs.piano.io/llms-full.txt
https://docs.piano.io/llms-full.txt/1
https://docs.piano.io/llms-full.txt/2
...and so on
Scoping by product or language
There are currently no dedicated per-product index files (e.g. llms-analytics.txt) on this site. Instead, scope your agent's crawl using the URL path prefix. Every page under a given product/locale shares a consistent prefix:
-
Piano Analytics: English:
docs.piano.io/en/analytics/, French:docs.piano.io/fr/analytics/ -
Piano Subscriptions: English:
docs.piano.io/en/subscriptions/ -
Piano Audience: English:
docs.piano.io/en/audience/ -
Piano Amplifier:
docs.piano.io/amplifier/
Tip for agents: fetch llms.txt once, then filter the returned links by the prefix that matches the product the user is asking about before fetching individual .md pages. This mimics scoped-index behavior without an extra request.
Per-page markdown files
Every documentation page also has a standalone .md export. Links in llms.txt point to these URLs directly.
Pattern: https://docs.piano.io/<route>.md
Example: en/analytics/create-modify-a-board.md
Quick start
Fetch an index or a single page from the command line:
curl -s https://docs.piano.io/llms.txt
curl -s https://docs.piano.io/llms-full.txt
curl -s https://docs.piano.io/en/analytics/create-modify-a-board.md
Example prompts
Copy and adapt these instructions for your AI tool of choice:
Cursor / Windsurf rule
When answering questions about Piano product documentation:
Fetch https://docs.piano.io/llms.txt first.
Filter the index for links matching the relevant product path (/en/analytics/, /en/subscriptions/, /en/audience/, /amplifier/).
Fetch only the matching .md pages, do not guess feature details.
For API/SDK reference questions, defer to developers.piano.io/analytics/llms.txt instead.
ChatGPT / Claude project instruction
You have access to Piano's product documentation via llms.txt exports at http://docs.piano.io . Start with https://docs.piano.io/llms.txt as your routing catalog. Identify the product the user is asking about (Analytics, Subscriptions, Audience, or Amplifier) and only fetch .md pages under that product's path prefix. Follow links to individual pages for authoritative, up-to-date answers.
RAG / agent workflow
Index step: ingest llms.txt as your routing catalog (small, cheap).
Retrieval step: fetch targeted .md URLs listed in the index. Avoid llms-full.txt unless you need broad, cross-product context, and expect to paginate through /1, /2, /3, etc.
Summaries in the index come from each page's title/description metadata.