feat(ts-sdk): add vLLM provider (#5805)

Co-authored-by: kartik-mem0 <kartik.labhshetwar@mem0.ai>

docs/components/llms/models/vllm.mdx

@@ -25,7 +25,8 @@ description: "Configure vLLM as an LLM provider in Mem0 for high-performance loc
 
 ## Usage
 
-```python
+<CodeGroup>
+```python Python
 import os
 from mem0 import Memory
 
@@ -53,6 +54,46 @@ messages = [
 m.add(messages, user_id="alice", metadata={"category": "movies"})
 ```
 
+```typescript TypeScript
+import { Memory } from "mem0ai/oss";
+
+const config = {
+  llm: {
+    provider: "vllm",
+    config: {
+      model: "Qwen/Qwen2.5-32B-Instruct",
+      baseURL: "http://localhost:8000/v1",
+      apiKey: process.env.VLLM_API_KEY || "vllm-api-key",
+      temperature: 0.1,
+      maxTokens: 2000,
+    },
+  },
+};
+
+const memory = new Memory(config);
+const messages = [
+  {
+    role: "user",
+    content: "I'm planning to watch a movie tonight. Any recommendations?",
+  },
+  {
+    role: "assistant",
+    content: "How about thriller movies? They can be quite engaging.",
+  },
+  {
+    role: "user",
+    content: "I'm not a big fan of thrillers, but I love sci-fi movies.",
+  },
+  {
+    role: "assistant",
+    content: "Got it! I'll avoid thrillers and suggest sci-fi movies instead.",
+  },
+];
+await memory.add(messages, { userId: "alice", metadata: { category: "movies" } });
+```
+
+</CodeGroup>
+
 ## Configuration Parameters
 
 | Parameter       | Description                       | Default                       | Environment Variable |

mem0-ts/src/oss/src/config/manager.ts

@@ -96,6 +96,8 @@ export class ConfigManager {
         config: (() => {
           const defaultConf = DEFAULT_MEMORY_CONFIG.llm.config;
           const userConf = userConfig.llm?.config;
+          const provider =
+            userConfig.llm?.provider || DEFAULT_MEMORY_CONFIG.llm.provider;
           let finalModel: string | any = defaultConf.model;
 
           if (userConf?.model && typeof userConf.model === "object") {
@@ -105,14 +107,18 @@ export class ConfigManager {
           }
 
           // Normalize snake_case keys from Python SDK / OpenClaw configs
+          const llmRaw = userConf as Record<string, unknown> | undefined;
           const llmBaseURL =
             userConf?.baseURL ??
+            userConf?.vllmBaseURL ??
+            (llmRaw?.vllm_base_url as string | undefined) ??
             ((userConf as Record<string, unknown>)?.lmstudio_base_url as
               | string
               | undefined) ??
             userConf?.url ??
-            defaultConf.baseURL;
-          const llmRaw = userConf as Record<string, unknown> | undefined;
+            (provider.toLowerCase() === "vllm"
+              ? undefined
+              : defaultConf.baseURL);
           const temperature =
             userConf?.temperature ??
             (llmRaw?.temperature as number | undefined);

mem0-ts/src/oss/src/index.ts

@@ -19,6 +19,7 @@ export * from "./llms/lmstudio";
 export * from "./llms/mistral";
 export * from "./llms/langchain";
 export * from "./llms/litellm";
+export * from "./llms/vllm";
 export * from "./vector_stores/base";
 export * from "./vector_stores/memory";
 export * from "./vector_stores/qdrant";

mem0-ts/src/oss/src/llms/vllm.ts

@@ -0,0 +1,49 @@
+import { OpenAILLM } from "./openai";
+import { LLMConfig, Message } from "../types";
+import { LLMResponse } from "./base";
+
+const DEFAULT_MODEL = "Qwen/Qwen2.5-32B-Instruct";
+const DEFAULT_API_KEY = "vllm-api-key";
+// Mirrors the Python provider default (mem0/configs/llms/vllm.py) and the docs table.
+const DEFAULT_BASE_URL = "http://localhost:8000/v1";
+
+export class VllmLLM extends OpenAILLM {
+  constructor(config: LLMConfig) {
+    const baseURL =
+      config.baseURL ||
+      config.vllmBaseURL ||
+      config.vllm_base_url ||
+      config.url ||
+      process.env.VLLM_BASE_URL ||
+      DEFAULT_BASE_URL;
+
+    super({
+      ...config,
+      apiKey: config.apiKey || process.env.VLLM_API_KEY || DEFAULT_API_KEY,
+      baseURL,
+      model: config.model || DEFAULT_MODEL,
+    });
+  }
+
+  async generateResponse(
+    messages: Message[],
+    responseFormat?: { type: string },
+    tools?: any[],
+  ): Promise<string | LLMResponse> {
+    try {
+      return await super.generateResponse(messages, responseFormat, tools);
+    } catch (err) {
+      const message = err instanceof Error ? err.message : String(err);
+      throw new Error(`vLLM LLM failed: ${message}`);
+    }
+  }
+
+  async generateChat(messages: Message[]): Promise<LLMResponse> {
+    try {
+      return await super.generateChat(messages);
+    } catch (err) {
+      const message = err instanceof Error ? err.message : String(err);
+      throw new Error(`vLLM LLM failed: ${message}`);
+    }
+  }
+}

mem0-ts/src/oss/src/tests/vllm.test.ts

@@ -0,0 +1,108 @@
+import OpenAI from "openai";
+import { VllmLLM } from "../llms/vllm";
+import { LLMFactory } from "../utils/factory";
+
+const createMock = jest.fn();
+
+jest.mock("openai", () => {
+  return jest.fn().mockImplementation(() => ({
+    chat: {
+      completions: {
+        create: createMock,
+      },
+    },
+  }));
+});
+
+const MockOpenAI = OpenAI as unknown as jest.Mock;
+
+describe("VllmLLM", () => {
+  beforeEach(() => {
+    jest.clearAllMocks();
+    delete process.env.VLLM_API_KEY;
+    delete process.env.VLLM_BASE_URL;
+    createMock.mockResolvedValue({
+      choices: [
+        {
+          message: {
+            content: "hello",
+            role: "assistant",
+          },
+        },
+      ],
+    });
+  });
+
+  it("uses vLLM defaults with a configured baseURL", async () => {
+    const llm = new VllmLLM({ baseURL: "http://localhost:8000/v1" });
+
+    await llm.generateChat([{ role: "user", content: "Hi" }]);
+
+    expect(MockOpenAI).toHaveBeenCalledWith({
+      apiKey: "vllm-api-key",
+      baseURL: "http://localhost:8000/v1",
+    });
+    expect(createMock).toHaveBeenCalledWith({
+      messages: [{ role: "user", content: "Hi" }],
+      model: "Qwen/Qwen2.5-32B-Instruct",
+    });
+  });
+
+  it("reads vLLM API settings from the environment", () => {
+    process.env.VLLM_API_KEY = "env-key";
+    process.env.VLLM_BASE_URL = "http://vllm.example/v1";
+
+    new VllmLLM({});
+
+    expect(MockOpenAI).toHaveBeenCalledWith({
+      apiKey: "env-key",
+      baseURL: "http://vllm.example/v1",
+    });
+  });
+
+  it("accepts Python-style vllm_base_url configs", () => {
+    new VllmLLM({ vllm_base_url: "http://localhost:8001/v1" });
+
+    expect(MockOpenAI).toHaveBeenCalledWith({
+      apiKey: "vllm-api-key",
+      baseURL: "http://localhost:8001/v1",
+    });
+  });
+
+  it("registers the vllm provider with LLMFactory", () => {
+    const llm = LLMFactory.create("vllm", {
+      baseURL: "http://localhost:8000/v1",
+    });
+
+    expect(llm).toBeInstanceOf(VllmLLM);
+  });
+
+  it("defaults to the local vLLM server when no baseURL is provided", () => {
+    new VllmLLM({});
+
+    expect(MockOpenAI).toHaveBeenCalledWith({
+      apiKey: "vllm-api-key",
+      baseURL: "http://localhost:8000/v1",
+    });
+  });
+
+  it("uses VLLM_BASE_URL when config merging leaves baseURL unset", () => {
+    process.env.VLLM_BASE_URL = "http://env-vllm.example/v1";
+
+    new VllmLLM({ model: "Qwen/Qwen2.5-32B-Instruct" });
+
+    expect(MockOpenAI).toHaveBeenCalledWith({
+      apiKey: "vllm-api-key",
+      baseURL: "http://env-vllm.example/v1",
+    });
+  });
+
+  it("wraps OpenAI-compatible client errors with provider context", async () => {
+    createMock.mockRejectedValueOnce(new Error("network down"));
+    const llm = new VllmLLM({ baseURL: "http://localhost:8000/v1" });
+
+    await expect(
+      llm.generateResponse([{ role: "user", content: "Hi" }]),
+    ).rejects.toThrow("vLLM LLM failed: network down");
+  });
+});

mem0-ts/src/oss/src/types/index.ts

@@ -45,6 +45,8 @@ export interface HistoryStoreConfig {
 export interface LLMConfig {
   provider?: string;
   baseURL?: string;
+  vllmBaseURL?: string;
+  vllm_base_url?: string;
   url?: string;
   config?: Record<string, any>;
   apiKey?: string;
@@ -135,6 +137,8 @@ export const MemoryConfigSchema = z.object({
       model: z.union([z.string(), z.any()]).optional(),
       modelProperties: z.record(z.string(), z.any()).optional(),
       baseURL: z.string().optional(),
+      vllmBaseURL: z.string().optional(),
+      vllm_base_url: z.string().optional(),
       url: z.string().optional(),
       timeout: z.number().optional(),
       temperature: z.number().optional(),

mem0-ts/src/oss/src/utils/factory.ts

@@ -25,6 +25,7 @@ import { LMStudioLLM } from "../llms/lmstudio";
 import { DeepSeekLLM } from "../llms/deepseek";
 import { LiteLLM } from "../llms/litellm";
 import { MiniMaxLLM } from "../llms/minimax";
+import { VllmLLM } from "../llms/vllm";
 import { SupabaseDB } from "../vector_stores/supabase";
 import { SQLiteManager } from "../storage/SQLiteManager";
 import { MemoryHistoryManager } from "../storage/MemoryHistoryManager";
@@ -93,6 +94,8 @@ export class LLMFactory {
         return new LiteLLM(config);
       case "minimax":
         return new MiniMaxLLM(config);
+      case "vllm":
+        return new VllmLLM(config);
       default:
         throw new Error(`Unsupported LLM provider: ${provider}`);
     }

mem0-ts/src/oss/tests/config-manager.test.ts

@@ -155,6 +155,35 @@ describe("ConfigManager", () => {
       expect(config.llm.config.baseURL).toBe("https://api.openai.com/v1");
     });
 
+    it("normalizes vllm_base_url to baseURL for vLLM", () => {
+      const config = ConfigManager.mergeConfig({
+        embedder: baseEmbedder,
+        vectorStore: baseVectorStore,
+        llm: {
+          provider: "vllm",
+          config: {
+            model: "Qwen/Qwen2.5-32B-Instruct",
+            vllm_base_url: "http://localhost:8000/v1",
+          },
+        },
+      });
+
+      expect(config.llm.config.baseURL).toBe("http://localhost:8000/v1");
+    });
+
+    it("does not inject the OpenAI baseURL default for vLLM", () => {
+      const config = ConfigManager.mergeConfig({
+        embedder: baseEmbedder,
+        vectorStore: baseVectorStore,
+        llm: {
+          provider: "vllm",
+          config: { model: "Qwen/Qwen2.5-32B-Instruct" },
+        },
+      });
+
+      expect(config.llm.config.baseURL).toBeUndefined();
+    });
+
     it("should preserve url in embedder config (existing behavior)", () => {
       const config = ConfigManager.mergeConfig({
         embedder: {

mem0-ts/src/oss/tests/factory.unit.test.ts

@@ -102,6 +102,11 @@ jest.mock("../src/llms/minimax", () => ({
     .fn()
     .mockImplementation((config) => ({ type: "minimax-llm", config })),
 }));
+jest.mock("../src/llms/vllm", () => ({
+  VllmLLM: jest
+    .fn()
+    .mockImplementation((config) => ({ type: "vllm-llm", config })),
+}));
 
 jest.mock("../src/vector_stores/qdrant", () => ({
   Qdrant: jest
@@ -228,6 +233,7 @@ describe("LLMFactory", () => {
     ["deepseek"],
     ["litellm"],
     ["minimax"],
+    ["vllm"],
   ])("creates LLM for provider '%s'", (provider) => {
     expect(() => LLMFactory.create(provider, dummyLLMConfig)).not.toThrow();
   });