Add Public AI (TEE) inference provider

packages/inference/README.md

@@ -55,6 +55,7 @@ Currently, we support the following providers:
 - [Nscale](https://nscale.com)
 - [OVHcloud](https://endpoints.ai.cloud.ovh.net/)
 - [Public AI](https://publicai.co)
+- [Public AI (TEE)](https://publicai.io)
 - [Replicate](https://replicate.com)
 - [Scaleway](https://www.scaleway.com/en/generative-apis/)
 - [Together](https://together.xyz)
@@ -90,6 +91,7 @@ Only a subset of models are supported when requesting third-party providers. You
 - [HF Inference supported models](https://huggingface.co/api/partners/hf-inference/models)
 - [Nscale supported models](https://huggingface.co/api/partners/nscale/models)
 - [OVHcloud supported models](https://huggingface.co/api/partners/ovhcloud/models)
+- [Public AI (TEE) supported models](https://huggingface.co/api/partners/publicai-tee/models)
 - [Replicate supported models](https://huggingface.co/api/partners/replicate/models)
 - [Scaleway supported models](https://huggingface.co/api/partners/scaleway/models)
 - [Together supported models](https://huggingface.co/api/partners/together/models)

packages/inference/src/lib/getProviderHelper.ts

@@ -12,6 +12,7 @@ import * as Nscale from "../providers/nscale.js";
 import * as OpenAI from "../providers/openai.js";
 import * as OvhCloud from "../providers/ovhcloud.js";
 import * as PublicAI from "../providers/publicai.js";
+import * as PublicAITee from "../providers/publicai-tee.js";
 import type {
 	AudioClassificationTaskHelper,
 	AudioToAudioTaskHelper,
@@ -143,6 +144,9 @@ export const PROVIDERS: Record<InferenceProvider, Partial<Record<InferenceTask,
 	publicai: {
 		conversational: new PublicAI.PublicAIConversationalTask(),
 	},
+	"publicai-tee": {
+		conversational: new PublicAITee.PublicAITeeConversationalTask(),
+	},
 	replicate: {
 		"text-to-image": new Replicate.ReplicateTextToImageTask(),
 		"text-to-speech": new Replicate.ReplicateTextToSpeechTask(),

packages/inference/src/providers/consts.ts

@@ -31,6 +31,7 @@ export const HARDCODED_MODEL_INFERENCE_MAPPING: Record<
 	nscale: {},
 	openai: {},
 	publicai: {},
+	"publicai-tee": {},
 	ovhcloud: {},
 	replicate: {},
 	scaleway: {},

packages/inference/src/providers/publicai-tee.ts

@@ -0,0 +1,7 @@
+import { BaseConversationalTask } from "./providerHelper.js";
+
+export class PublicAITeeConversationalTask extends BaseConversationalTask {
+	constructor() {
+		super("publicai-tee", "https://tee.publicai.io");
+	}
+}

packages/inference/src/types.ts

@@ -59,6 +59,7 @@ export const INFERENCE_PROVIDERS = [
 	"openai",
 	"ovhcloud",
 	"publicai",
+	"publicai-tee",
 	"replicate",
 	"scaleway",
 	"together",
@@ -92,6 +93,7 @@ export const PROVIDERS_HUB_ORGS: Record<InferenceProvider, string> = {
 	openai: "openai",
 	ovhcloud: "ovhcloud",
 	publicai: "publicai",
+	"publicai-tee": "publicai-tee",
 	replicate: "replicate",
 	scaleway: "scaleway",
 	together: "togethercomputer",

packages/inference/test/InferenceClient.spec.ts

@@ -2325,6 +2325,59 @@ describe.skip("InferenceClient", () => {
 		TIMEOUT,
 	);
 
+	describe.concurrent(
+		"PublicAI TEE",
+		() => {
+			const client = new InferenceClient(env.HF_PUBLICAI_TEE_KEY ?? "dummy");
+
+			HARDCODED_MODEL_INFERENCE_MAPPING["publicai-tee"] = {
+				"MiniMaxAI/MiniMax-M2": {
+					provider: "publicai-tee",
+					hfModelId: "MiniMaxAI/MiniMax-M2",
+					providerId: "MiniMax-M2",
+					status: "live",
+					task: "conversational",
+				},
+			};
+
+			it("chatCompletion", async () => {
+				const res = await client.chatCompletion({
+					model: "MiniMaxAI/MiniMax-M2",
+					provider: "publicai-tee",
+					messages: [{ role: "user", content: "Complete this sentence with words, one plus one is equal " }],
+				});
+				if (res.choices && res.choices.length > 0) {
+					const completion = res.choices[0].message?.content;
+					expect(completion).toContain("two");
+				}
+			});
+
+			it("chatCompletion stream", async () => {
+				const stream = client.chatCompletionStream({
+					model: "MiniMaxAI/MiniMax-M2",
+					provider: "publicai-tee",
+					messages: [{ role: "user", content: "Say 'this is a test'" }],
+					stream: true,
+				}) as AsyncGenerator<ChatCompletionStreamOutput>;
+
+				let fullResponse = "";
+				for await (const chunk of stream) {
+					if (chunk.choices && chunk.choices.length > 0) {
+						const content = chunk.choices[0].delta?.content;
+						if (content) {
+							fullResponse += content;
+						}
+					}
+				}
+
+				// Verify we got a meaningful response
+				expect(fullResponse).toBeTruthy();
+				expect(fullResponse.length).toBeGreaterThan(0);
+			});
+		},
+		TIMEOUT,
+	);
+
 	describe.concurrent(
 		"Baseten",
 		() => {