Rate Limits & Caching

There is no enforced rate limit on the public Doungim REST API or MCP server, but a fair-use policy applies and high-volume callers are asked to cache aggressively. Doungim is a TTRPG gaming console for D&D and other tabletop role-playing games; the public API exists to spread the data, not to absorb sustained heavy traffic.

CDN cache headers

Every REST response carries:

Cache-Control: public, max-age=300, s-maxage=3600, stale-while-revalidate=86400

That's 5 minutes in the browser, 1 hour at the edge, and a day of stale-while-revalidate. The underlying data updates rarely — most weeks not at all — so this is conservative.

Recommended client-side TTLs

• Static spells / monsters / items: 24 hours, refresh on app launch.
• Search results: 1 hour (queries are bounded, the index is stable).
• Encounter tables: 24 hours.

Mirror the data

Apps that ship to users (mobile, desktop, offline-first web) should pull the whole list endpoints once at build time or once a week:

# Pull every spell once
curl https://www.doungim.com/api/v1/dnd/spells > spells.json

# Pull every monster index
curl https://www.doungim.com/api/v1/dnd/monsters > monsters.json

Then fetch individual /{slug} entries only on demand and cache them.

Identify yourself

Please set a recognisable User-Agent so we can spot a misbehaving client without blocking your whole IP range:

User-Agent: MyApp/1.2 (contact: alice@example.com)

Abuse policy

We may rate-limit or block clients that:

• Hammer endpoints faster than 10 requests per second sustained.
• Skip caching and re-fetch the same page on every render.
• Strip the citation block from responses before redisplaying the data.