feat(providers): add DigitalOcean AI as an OpenAI-compatible provider

DigitalOcean AI serverless inference (base https://inference.do-ai.run/v1)
registered as an API-key OpenAI-compatible provider with full model
passthrough. Wired through the shared OpenAI-compatible registry
(inference-hosts family). No new deps, no schema change.

Co-authored-by: newnol <tantai13102005@gmail.com>
Inspired-by: https://github.com/decolua/9router/pull/2417

src/shared/constants/providers/apikey/inference-hosts.ts

@@ -309,4 +309,20 @@ export const APIKEY_PROVIDERS_INFERENCE = {
     },
     serviceKinds: ["llm"],
   },
+  digitalocean: {
+    id: "digitalocean",
+    alias: "digitalocean",
+    name: "DigitalOcean",
+    icon: "cloud",
+    color: "#0060FF",
+    textIcon: "DO",
+    passthroughModels: true,
+    website: "https://docs.digitalocean.com/products/ai-platform/",
+    hasFree: false,
+    notice: {
+      text: "Use a DigitalOcean Personal Access Token (dop_v1_...) or a Model Access Key from the Inference console. OAuth tokens (doo_v1_...) may not have the required scopes.",
+      apiKeyUrl: "https://cloud.digitalocean.com/account/api/tokens",
+    },
+    serviceKinds: ["llm"],
+  },
 };