For AI agents & LLMs

The Crius documentation is built to be read by language models as easily as by people.

Resource	URL	Use
llms.txt	https://crius.sh/llms.txt	Index of the docs with one-line summaries (llmstxt.org format)
llms-full.txt	https://crius.sh/llms-full.txt	All documentation in one Markdown file
Markdown pages	https://crius.sh/docs/<page>.md	Any docs page as clean Markdown, e.g. /docs/quickstart.md
OpenAPI 3	https://api.crius.sh/openapi.json	Machine-readable endpoints and parameters, for tool or function definitions

Every docs page also has a Copy page button that copies its Markdown, ready to paste into a chat.

Giving an agent access to the API

• Core endpoints (prices, token profiles, pools, trades, candles, search, stats) work without a key, so an agent can explore at 2 requests per second.
• For anything more, create a dedicated API key for the agent with only the read scope, an IP allow-list if it runs on a fixed host, and a per-key rate limit. Rotate or revoke it from the dashboard at any time.
• Responses are compact JSON with plain decimal numbers and a meta.block that lets the agent cite exactly which block its answer reflects.

Suggested system prompt snippet

You can call the Crius API (https://api.crius.sh) for Robinhood Chain (chain id 4663) data.
Docs: https://crius.sh/llms.txt. Send the key in the x-api-key header.
Token price: GET /v1/robinhood/token/{address}/price. Token profile: GET /v1/robinhood/token/{address}.
Candles: GET /v1/robinhood/token/{address}/candles?interval=1h. Search: GET /v1/robinhood/search?q=.
Amounts are in human units, USD values are plain decimals, timestamps are unix milliseconds.