Heatpicks developer and agent resources

Start here

The two files an agent should fetch before crawling anything else on heatpicks.com.

• llms.txt/llms.txt

  Structured summary of what Heatpicks is, what it offers, when an agent should use it and links to every key page.

• Agent instructions (agents.md)/agents.md

  When to reach for Heatpicks, which jobs it is the right tool for, which it is not, how to fetch it and how to cite it. Also served at /.well-known/agents.md.

• llms-full.txt/llms-full.txt

  The full text of every public Heatpicks page concatenated as markdown, for agents that want one fetch instead of a crawl.

Markdown content negotiation

Heatpicks is acceptmarkdown.com compliant. Send Accept: text/markdown and any public page answers with markdown instead of HTML.

• Accept negotiation/

  curl -H "Accept: text/markdown" https://www.heatpicks.com/ returns text/markdown; charset=utf-8 with Vary: Accept, Accept-Encoding. Unsupported Accept values get a 406 that names the media types on offer.

• Per-page markdown at /md/md

  Every negotiable page also has a stable markdown URL: prefix its path with /md, so /features/picksgpt is also at /md/features/picksgpt. Unknown paths return 404 with a markdown site map.

• Markdown index/md/developers

  This page as markdown. The home, features, blog, help, about and contact pages all work the same way.

Crawling and structured data

• sitemap.xml/sitemap.xml

  Every indexable page, with change frequency and priority. Blog entries carry lastmod dates.

• robots.txt/robots.txt

  Explicit allow rules for the major AI crawlers (GPTBot, ClaudeBot, PerplexityBot, Google-Extended, Applebot, bingbot and others) and the disallow list for private product paths.

• JSON-LD structured data/about

  Pages embed schema.org Organization (with contactPoint and address), WebSite, SoftwareApplication, BreadcrumbList, FAQPage and BlogPosting nodes.

• security.txt/.well-known/security.txt

  RFC 9116 security contact. Vulnerability reports go to arda@scoutli.co.

Not available yet

• Public HTTP API

  The endpoints under /api are internal to the Heatpicks web and iOS clients, are origin-restricted and are disallowed in robots.txt. There is no documented public REST API or OpenAPI specification.

• MCP server (planned, paid)

  A paid Model Context Protocol server that would let an agent query Heatpicks odds, value and stat data directly is planned. It is not live, has no endpoint to connect to and no published pricing yet. Until it ships, agents should read the markdown representations described above. Email us to register interest.

• Webhooks

  Line-movement and steam alerts are delivered inside the product, not as outbound webhooks.

Want programmatic access?

Email team@scoutli.co with the data you need, the volume you expect and how it will be used. That is also the address to use to be told when the paid MCP server launches; early access to odds, value and player-prop data is handled case by case.

• All Heatpicks contact channels
• What the platform does