# Developer and Agent Resources

BoxOfChocolates.ai provides public, read-only REST and MCP interfaces for official organization and initiative information. They expose no private records, user accounts, write operations, medical uploads, GuardDog searches, or automated decisions.

## REST API

- Base URL: [`/api/v1`](https://boxofchocolates.ai/api/v1)
- OpenAPI 3.1: [`/openapi.json`](https://boxofchocolates.ai/openapi.json)
- Authentication: none
- Operations: organization details, initiative list, and initiative lookup
- Lifecycle and versioning: [`/api-versioning.md`](https://boxofchocolates.ai/api-versioning.md); no `v1` deprecation or sunset is currently scheduled

## MCP

- Streamable HTTP endpoint: [`/mcp`](https://boxofchocolates.ai/mcp)
- Discovery endpoint: [`/.well-known/mcp`](https://boxofchocolates.ai/.well-known/mcp)
- Server card: [`/.well-known/mcp/server-card.json`](https://boxofchocolates.ai/.well-known/mcp/server-card.json)
- Authentication: none
- Tools: `get_organization`, `list_initiatives`, and `get_initiative`
- Resources: organization and initiative JSON

## Machine-readable resources

- [`/llms.txt`](https://boxofchocolates.ai/llms.txt) — concise agent navigation and use guidance
- [`/llms-full.txt`](https://boxofchocolates.ai/llms-full.txt) — fuller organizational context
- [`/index.md`](https://boxofchocolates.ai/index.md) — Markdown representation of the homepage
- [`/sitemap.xml`](https://boxofchocolates.ai/sitemap.xml) — indexable public pages
- [`/docs/llms.txt`](https://boxofchocolates.ai/docs/llms.txt) — scoped API and MCP index
- [`/.well-known/ai-catalog.json`](https://boxofchocolates.ai/.well-known/ai-catalog.json) — Agentic Resource Discovery catalog
- [`/.well-known/api-catalog`](https://boxofchocolates.ai/.well-known/api-catalog) — RFC 9727 API catalog
- [`/.well-known/agent-skills/index.json`](https://boxofchocolates.ai/.well-known/agent-skills/index.json) — Agent Skills discovery index

## Safety and versioning

The interfaces are versioned at `v1` and return structured JSON errors. Initiative results include explicit safety boundaries. They are informational only: fcancer.ai is not diagnosis, and GuardDog is not a blacklist, consumer background check, or automatic adoption-decision engine.