Add new model provider PPIO (#3211)

* feat: add new model provider PPIO

* fix: fix ppio model fetching

* fix: code lint

* reorder LLM
update interface for streaming and chats to use valid keys
linting

---------

Co-authored-by: timothycarambat <rambat1010@gmail.com>

.vscode/settings.json

@@ -41,6 +41,7 @@
     "opendocument",
     "openrouter",
     "pagerender",
+    "ppio",
     "Qdrant",
     "royalblue",
     "SearchApi",

README.md

@@ -99,6 +99,7 @@ AnythingLLM divides your documents into objects called `workspaces`. A Workspace
 - [Apipie](https://apipie.ai/)
 - [xAI](https://x.ai/)
 - [Novita AI (chat models)](https://novita.ai/model-api/product/llm-api?utm_source=github_anything-llm&utm_medium=github_readme&utm_campaign=link)
+- [PPIO](https://ppinfra.com?utm_source=github_anything-llm)
 
 **Embedder models:**
 

docker/.env.example

@@ -126,6 +126,10 @@ GID='1000'
 # DEEPSEEK_API_KEY='your-deepseek-api-key-here'
 # DEEPSEEK_MODEL_PREF='deepseek-chat'
 
+# LLM_PROVIDER='ppio'
+# PPIO_API_KEY='your-ppio-api-key-here'
+# PPIO_MODEL_PREF=deepseek/deepseek-v3/community
+
 ###########################################
 ######## Embedding API SElECTION ##########
 ###########################################

frontend/src/components/LLMSelection/PPIOLLMOptions/index.jsx

@@ -0,0 +1,100 @@
+import System from "@/models/system";
+import { useState, useEffect } from "react";
+
+export default function PPIOLLMOptions({ settings }) {
+  return (
+    <div className="w-full flex flex-col gap-y-7">
+      <div className="w-full flex items-start gap-[36px] mt-1.5">
+        <div className="flex flex-col w-60">
+          <label className="text-theme-text-primary text-sm font-semibold block mb-3">
+            PPIO API Key
+          </label>
+          <input
+            type="password"
+            name="PPIOApiKey"
+            className="border-none bg-theme-settings-input-bg text-theme-text-primary placeholder:text-theme-settings-input-placeholder text-sm rounded-lg focus:outline-primary-button active:outline-primary-button outline-none block w-full p-2.5"
+            placeholder="PPIO API Key"
+            defaultValue={settings?.PPIOApiKey ? "*".repeat(20) : ""}
+            required={true}
+            autoComplete="off"
+            spellCheck={false}
+          />
+        </div>
+        {!settings?.credentialsOnly && (
+          <PPIOModelSelection settings={settings} />
+        )}
+      </div>
+    </div>
+  );
+}
+
+function PPIOModelSelection({ settings }) {
+  const [groupedModels, setGroupedModels] = useState({});
+  const [loading, setLoading] = useState(true);
+
+  useEffect(() => {
+    async function fetchModels() {
+      setLoading(true);
+      const { models } = await System.customModels("ppio");
+      if (models?.length > 0) {
+        const modelsByOrganization = models.reduce((acc, model) => {
+          acc[model.organization] = acc[model.organization] || [];
+          acc[model.organization].push(model);
+          return acc;
+        }, {});
+        setGroupedModels(modelsByOrganization);
+      }
+      setLoading(false);
+    }
+    fetchModels();
+  }, []);
+
+  if (loading || Object.keys(groupedModels).length === 0) {
+    return (
+      <div className="flex flex-col w-60">
+        <label className="text-theme-text-primary text-sm font-semibold block mb-3">
+          Chat Model Selection
+        </label>
+        <select
+          name="PPIOModelPref"
+          required={true}
+          disabled={true}
+          className="bg-theme-settings-input-bg text-theme-text-primary text-sm rounded-lg focus:ring-primary-button focus:border-primary-button block w-full p-2.5"
+        >
+          <option disabled={true} selected={true}>
+            -- loading available models --
+          </option>
+        </select>
+      </div>
+    );
+  }
+
+  return (
+    <div className="flex flex-col">
+      <label className="text-theme-text-primary text-sm font-semibold block mb-3">
+        Chat Model Selection
+      </label>
+      <select
+        name="PPIOModelPref"
+        required={true}
+        className="border-none bg-theme-settings-input-bg text-theme-text-primary border-theme-border text-sm rounded-lg block w-full p-2.5"
+      >
+        {Object.keys(groupedModels)
+          .sort()
+          .map((organization) => (
+            <optgroup key={organization} label={organization}>
+              {groupedModels[organization].map((model) => (
+                <option
+                  key={model.id}
+                  value={model.id}
+                  selected={settings?.PPIOModelPref === model.id}
+                >
+                  {model.name}
+                </option>
+              ))}
+            </optgroup>
+          ))}
+      </select>
+    </div>
+  );
+}

frontend/src/hooks/useGetProvidersModels.js

@@ -79,6 +79,7 @@ const groupedProviders = [
   "openai",
   "novita",
   "openrouter",
+  "ppio",
 ];
 export default function useGetProviderModels(provider = null) {
   const [defaultModels, setDefaultModels] = useState([]);

frontend/src/pages/GeneralSettings/LLMPreference/index.jsx

@@ -30,6 +30,7 @@ import DeepSeekLogo from "@/media/llmprovider/deepseek.png";
 import APIPieLogo from "@/media/llmprovider/apipie.png";
 import XAILogo from "@/media/llmprovider/xai.png";
 import NvidiaNimLogo from "@/media/llmprovider/nvidia-nim.png";
+import PPIOLogo from "@/media/llmprovider/ppio.png";
 
 import PreLoader from "@/components/Preloader";
 import OpenAiOptions from "@/components/LLMSelection/OpenAiOptions";
@@ -57,6 +58,7 @@ import DeepSeekOptions from "@/components/LLMSelection/DeepSeekOptions";
 import ApiPieLLMOptions from "@/components/LLMSelection/ApiPieOptions";
 import XAILLMOptions from "@/components/LLMSelection/XAiLLMOptions";
 import NvidiaNimOptions from "@/components/LLMSelection/NvidiaNimOptions";
+import PPIOLLMOptions from "@/components/LLMSelection/PPIOLLMOptions";
 
 import LLMItem from "@/components/LLMSelection/LLMItem";
 import { CaretUpDown, MagnifyingGlass, X } from "@phosphor-icons/react";
@@ -246,6 +248,15 @@ export const AVAILABLE_LLM_PROVIDERS = [
     description: "Run DeepSeek's powerful LLMs.",
     requiredConfig: ["DeepSeekApiKey"],
   },
+  {
+    name: "PPIO",
+    value: "ppio",
+    logo: PPIOLogo,
+    options: (settings) => <PPIOLLMOptions settings={settings} />,
+    description:
+      "Run stable and cost-efficient open-source LLM APIs, such as DeepSeek, Llama, Qwen etc.",
+    requiredConfig: ["PPIOApiKey"],
+  },
   {
     name: "AWS Bedrock",
     value: "bedrock",

frontend/src/pages/OnboardingFlow/Steps/DataHandling/index.jsx

@@ -25,7 +25,6 @@ import AWSBedrockLogo from "@/media/llmprovider/bedrock.png";
 import DeepSeekLogo from "@/media/llmprovider/deepseek.png";
 import APIPieLogo from "@/media/llmprovider/apipie.png";
 import XAILogo from "@/media/llmprovider/xai.png";
-
 import CohereLogo from "@/media/llmprovider/cohere.png";
 import ZillizLogo from "@/media/vectordbs/zilliz.png";
 import AstraDBLogo from "@/media/vectordbs/astraDB.png";
@@ -36,6 +35,7 @@ import WeaviateLogo from "@/media/vectordbs/weaviate.png";
 import QDrantLogo from "@/media/vectordbs/qdrant.png";
 import MilvusLogo from "@/media/vectordbs/milvus.png";
 import VoyageAiLogo from "@/media/embeddingprovider/voyageai.png";
+import PPIOLogo from "@/media/llmprovider/ppio.png";
 
 import React, { useState, useEffect } from "react";
 import paths from "@/utils/paths";
@@ -226,6 +226,14 @@ export const LLM_SELECTION_PRIVACY = {
     ],
     logo: XAILogo,
   },
+  ppio: {
+    name: "PPIO",
+    description: [
+      "Your chats will not be used for training",
+      "Your prompts and document text used in response creation are visible to PPIO",
+    ],
+    logo: PPIOLogo,
+  },
 };
 
 export const VECTOR_DB_PRIVACY = {

frontend/src/pages/OnboardingFlow/Steps/LLMPreference/index.jsx

@@ -25,6 +25,8 @@ import NovitaLogo from "@/media/llmprovider/novita.png";
 import XAILogo from "@/media/llmprovider/xai.png";
 import NvidiaNimLogo from "@/media/llmprovider/nvidia-nim.png";
 import CohereLogo from "@/media/llmprovider/cohere.png";
+import PPIOLogo from "@/media/llmprovider/ppio.png";
+
 import OpenAiOptions from "@/components/LLMSelection/OpenAiOptions";
 import GenericOpenAiOptions from "@/components/LLMSelection/GenericOpenAiOptions";
 import AzureAiOptions from "@/components/LLMSelection/AzureAiOptions";
@@ -50,6 +52,7 @@ import ApiPieLLMOptions from "@/components/LLMSelection/ApiPieOptions";
 import NovitaLLMOptions from "@/components/LLMSelection/NovitaLLMOptions";
 import XAILLMOptions from "@/components/LLMSelection/XAiLLMOptions";
 import NvidiaNimOptions from "@/components/LLMSelection/NvidiaNimOptions";
+import PPIOLLMOptions from "@/components/LLMSelection/PPIOLLMOptions";
 
 import LLMItem from "@/components/LLMSelection/LLMItem";
 import System from "@/models/system";
@@ -213,6 +216,14 @@ const LLMS = [
     options: (settings) => <DeepSeekOptions settings={settings} />,
     description: "Run DeepSeek's powerful LLMs.",
   },
+  {
+    name: "PPIO",
+    value: "ppio",
+    logo: PPIOLogo,
+    options: (settings) => <PPIOLLMOptions settings={settings} />,
+    description:
+      "Run stable and cost-efficient open-source LLM APIs, such as DeepSeek, Llama, Qwen etc.",
+  },
   {
     name: "APIpie",
     value: "apipie",

frontend/src/pages/WorkspaceSettings/AgentConfig/AgentLLMSelection/index.jsx

@@ -25,6 +25,7 @@ const ENABLED_PROVIDERS = [
   "bedrock",
   "fireworksai",
   "deepseek",
+  "ppio",
   "litellm",
   "apipie",
   "xai",

locales/README.fa-IR.md

@@ -103,6 +103,7 @@ AnythingLLM اسناد شما را به اشیایی به نام `workspaces` ت
 - [Apipie](https://apipie.ai/)
 - [xAI](https://x.ai/)
 - [Novita AI (chat models)](https://novita.ai/model-api/product/llm-api?utm_source=github_anything-llm&utm_medium=github_readme&utm_campaign=link)
+- [PPIO](https://ppinfra.com?utm_source=github_anything-llm)
 
 <div dir="rtl">
 

locales/README.ja-JP.md

@@ -90,6 +90,7 @@ AnythingLLMのいくつかのクールな機能
 - [Groq](https://groq.com/)
 - [Cohere](https://cohere.com/)
 - [KoboldCPP](https://github.com/LostRuins/koboldcpp)
+- [PPIO](https://ppinfra.com?utm_source=github_anything-llm)
 
 **埋め込みモデル：**
 

locales/README.tr-TR.md

@@ -100,6 +100,7 @@ AnythingLLM, belgelerinizi **"çalışma alanları" (workspaces)** adı verilen
 - [Apipie](https://apipie.ai/)
 - [xAI](https://x.ai/)
 - [Novita AI (chat models)](https://novita.ai/model-api/product/llm-api?utm_source=github_anything-llm&utm_medium=github_readme&utm_campaign=link)
+- [PPIO](https://ppinfra.com?utm_source=github_anything-llm)
 
 **Embedder modelleri:**
 

locales/README.zh-CN.md

@@ -86,6 +86,7 @@ AnythingLLM的一些酷炫特性
 - [Groq](https://groq.com/)
 - [Cohere](https://cohere.com/)
 - [KoboldCPP](https://github.com/LostRuins/koboldcpp)
+- [PPIO (聊天模型)](https://ppinfra.com?utm_source=github_anything-llm)
 
 **支持的嵌入模型：**
 

server/.env.example

@@ -116,6 +116,10 @@ SIG_SALT='salt' # Please generate random string at least 32 chars long.
 # NVIDIA_NIM_LLM_BASE_PATH='http://127.0.0.1:8000'
 # NVIDIA_NIM_LLM_MODEL_PREF='meta/llama-3.2-3b-instruct'
 
+# LLM_PROVIDER='ppio'
+# PPIO_API_KEY='your-ppio-api-key-here'
+# PPIO_MODEL_PREF='deepseek/deepseek-v3/community'
+
 ###########################################
 ######## Embedding API SElECTION ##########
 ###########################################

server/models/systemSettings.js

@@ -560,6 +560,10 @@ const SystemSettings = {
       NvidiaNimLLMBasePath: process.env.NVIDIA_NIM_LLM_BASE_PATH,
       NvidiaNimLLMModelPref: process.env.NVIDIA_NIM_LLM_MODEL_PREF,
       NvidiaNimLLMTokenLimit: process.env.NVIDIA_NIM_LLM_MODEL_TOKEN_LIMIT,
+
+      // PPIO API keys
+      PPIOApiKey: !!process.env.PPIO_API_KEY,
+      PPIOModelPref: process.env.PPIO_MODEL_PREF,
     };
   },
 

server/storage/models/.gitignore

@@ -8,4 +8,4 @@ mixedbread-ai*
 gemini
 togetherAi
 tesseract
-ppio
+ppio

server/utils/AiProviders/ppio/index.js

@@ -0,0 +1,266 @@
+const { NativeEmbedder } = require("../../EmbeddingEngines/native");
+const {
+  handleDefaultStreamResponseV2,
+} = require("../../helpers/chat/responses");
+const fs = require("fs");
+const path = require("path");
+const { safeJsonParse } = require("../../http");
+const {
+  LLMPerformanceMonitor,
+} = require("../../helpers/chat/LLMPerformanceMonitor");
+const cacheFolder = path.resolve(
+  process.env.STORAGE_DIR
+    ? path.resolve(process.env.STORAGE_DIR, "models", "ppio")
+    : path.resolve(__dirname, `../../../storage/models/ppio`)
+);
+
+class PPIOLLM {
+  constructor(embedder = null, modelPreference = null) {
+    if (!process.env.PPIO_API_KEY) throw new Error("No PPIO API key was set.");
+
+    const { OpenAI: OpenAIApi } = require("openai");
+    this.basePath = "https://api.ppinfra.com/v3/openai/";
+    this.openai = new OpenAIApi({
+      baseURL: this.basePath,
+      apiKey: process.env.PPIO_API_KEY ?? null,
+      defaultHeaders: {
+        "HTTP-Referer": "https://anythingllm.com",
+        "X-API-Source": "anythingllm",
+      },
+    });
+    this.model =
+      modelPreference ||
+      process.env.PPIO_MODEL_PREF ||
+      "qwen/qwen2.5-32b-instruct";
+    this.limits = {
+      history: this.promptWindowLimit() * 0.15,
+      system: this.promptWindowLimit() * 0.15,
+      user: this.promptWindowLimit() * 0.7,
+    };
+
+    this.embedder = embedder ?? new NativeEmbedder();
+    this.defaultTemp = 0.7;
+
+    if (!fs.existsSync(cacheFolder))
+      fs.mkdirSync(cacheFolder, { recursive: true });
+    this.cacheModelPath = path.resolve(cacheFolder, "models.json");
+    this.cacheAtPath = path.resolve(cacheFolder, ".cached_at");
+
+    this.log(`Loaded with model: ${this.model}`);
+  }
+
+  log(text, ...args) {
+    console.log(`\x1b[36m[${this.constructor.name}]\x1b[0m ${text}`, ...args);
+  }
+
+  async #syncModels() {
+    if (fs.existsSync(this.cacheModelPath) && !this.#cacheIsStale())
+      return false;
+
+    this.log("Model cache is not present or stale. Fetching from PPIO API.");
+    await fetchPPIOModels();
+    return;
+  }
+
+  #cacheIsStale() {
+    const MAX_STALE = 6.048e8; // 1 Week in MS
+    if (!fs.existsSync(this.cacheAtPath)) return true;
+    const now = Number(new Date());
+    const timestampMs = Number(fs.readFileSync(this.cacheAtPath));
+    return now - timestampMs > MAX_STALE;
+  }
+
+  #appendContext(contextTexts = []) {
+    if (!contextTexts || !contextTexts.length) return "";
+    return (
+      "\nContext:\n" +
+      contextTexts
+        .map((text, i) => {
+          return `[CONTEXT ${i}]:\n${text}\n[END CONTEXT ${i}]\n\n`;
+        })
+        .join("")
+    );
+  }
+
+  models() {
+    if (!fs.existsSync(this.cacheModelPath)) return {};
+    return safeJsonParse(
+      fs.readFileSync(this.cacheModelPath, { encoding: "utf-8" }),
+      {}
+    );
+  }
+
+  streamingEnabled() {
+    return "streamGetChatCompletion" in this;
+  }
+
+  promptWindowLimit() {
+    const model = this.models()[this.model];
+    if (!model) return 4096; // Default to 4096 if we cannot find the model
+    return model?.maxLength || 4096;
+  }
+
+  async isValidChatCompletionModel(model = "") {
+    await this.#syncModels();
+    const availableModels = this.models();
+    return Object.prototype.hasOwnProperty.call(availableModels, model);
+  }
+
+  /**
+   * Generates appropriate content array for a message + attachments.
+   * @param {{userPrompt:string, attachments: import("../../helpers").Attachment[]}}
+   * @returns {string|object[]}
+   */
+  #generateContent({ userPrompt, attachments = [] }) {
+    if (!attachments.length) {
+      return userPrompt;
+    }
+
+    const content = [{ type: "text", text: userPrompt }];
+    for (let attachment of attachments) {
+      content.push({
+        type: "image_url",
+        image_url: {
+          url: attachment.contentString,
+          detail: "auto",
+        },
+      });
+    }
+    return content.flat();
+  }
+
+  constructPrompt({
+    systemPrompt = "",
+    contextTexts = [],
+    chatHistory = [],
+    userPrompt = "",
+    // attachments = [], - not supported
+  }) {
+    const prompt = {
+      role: "system",
+      content: `${systemPrompt}${this.#appendContext(contextTexts)}`,
+    };
+    return [prompt, ...chatHistory, { role: "user", content: userPrompt }];
+  }
+
+  async getChatCompletion(messages = null, { temperature = 0.7 }) {
+    if (!(await this.isValidChatCompletionModel(this.model)))
+      throw new Error(
+        `PPIO chat: ${this.model} is not valid for chat completion!`
+      );
+
+    const result = await LLMPerformanceMonitor.measureAsyncFunction(
+      this.openai.chat.completions
+        .create({
+          model: this.model,
+          messages,
+          temperature,
+        })
+        .catch((e) => {
+          throw new Error(e.message);
+        })
+    );
+
+    if (
+      !Object.prototype.hasOwnProperty.call(result.output, "choices") ||
+      result.output.choices.length === 0
+    )
+      return null;
+
+    return {
+      textResponse: result.output.choices[0].message.content,
+      metrics: {
+        prompt_tokens: result.output.usage.prompt_tokens || 0,
+        completion_tokens: result.output.usage.completion_tokens || 0,
+        total_tokens: result.output.usage.total_tokens || 0,
+        outputTps: result.output.usage.completion_tokens / result.duration,
+        duration: result.duration,
+      },
+    };
+  }
+
+  async streamGetChatCompletion(messages = null, { temperature = 0.7 }) {
+    if (!(await this.isValidChatCompletionModel(this.model)))
+      throw new Error(
+        `PPIO chat: ${this.model} is not valid for chat completion!`
+      );
+
+    const measuredStreamRequest = await LLMPerformanceMonitor.measureStream(
+      this.openai.chat.completions.create({
+        model: this.model,
+        stream: true,
+        messages,
+        temperature,
+      }),
+      messages
+    );
+    return measuredStreamRequest;
+  }
+
+  handleStream(response, stream, responseProps) {
+    return handleDefaultStreamResponseV2(response, stream, responseProps);
+  }
+
+  async embedTextInput(textInput) {
+    return await this.embedder.embedTextInput(textInput);
+  }
+  async embedChunks(textChunks = []) {
+    return await this.embedder.embedChunks(textChunks);
+  }
+
+  async compressMessages(promptArgs = {}, rawHistory = []) {
+    const { messageArrayCompressor } = require("../../helpers/chat");
+    const messageArray = this.constructPrompt(promptArgs);
+    return await messageArrayCompressor(this, messageArray, rawHistory);
+  }
+}
+
+async function fetchPPIOModels() {
+  return await fetch(`https://api.ppinfra.com/v3/openai/models`, {
+    method: "GET",
+    headers: {
+      "Content-Type": "application/json",
+      Authorization: `Bearer ${process.env.PPIO_API_KEY}`,
+    },
+  })
+    .then((res) => res.json())
+    .then(({ data = [] }) => {
+      const models = {};
+      data.forEach((model) => {
+        const organization = model.id?.split("/")?.[0] || "PPIO";
+        models[model.id] = {
+          id: model.id,
+          name: model.display_name || model.title || model.id,
+          organization,
+          maxLength: model.context_size || 4096,
+        };
+      });
+
+      if (!fs.existsSync(cacheFolder))
+        fs.mkdirSync(cacheFolder, { recursive: true });
+      fs.writeFileSync(
+        path.resolve(cacheFolder, "models.json"),
+        JSON.stringify(models),
+        {
+          encoding: "utf-8",
+        }
+      );
+      fs.writeFileSync(
+        path.resolve(cacheFolder, ".cached_at"),
+        String(Number(new Date())),
+        {
+          encoding: "utf-8",
+        }
+      );
+      return models;
+    })
+    .catch((e) => {
+      console.error(e);
+      return {};
+    });
+}
+
+module.exports = {
+  PPIOLLM,
+  fetchPPIOModels,
+};

server/utils/agents/aibitat/index.js

@@ -795,7 +795,8 @@ ${this.getHistory({ to: route.to })
         return new Providers.XAIProvider({ model: config.model });
       case "novita":
         return new Providers.NovitaProvider({ model: config.model });
-
+      case "ppio":
+        return new Providers.PPIOProvider({ model: config.model });
       default:
         throw new Error(
           `Unknown provider: ${config.provider}. Please use a valid provider.`

server/utils/agents/aibitat/providers/ai-provider.js

@@ -163,6 +163,14 @@ class Provider {
           apiKey: process.env.NOVITA_LLM_API_KEY ?? null,
           ...config,
         });
+      case "ppio":
+        return new ChatOpenAI({
+          configuration: {
+            baseURL: "https://api.ppinfra.com/v3/openai",
+          },
+          apiKey: process.env.PPIO_API_KEY ?? null,
+          ...config,
+        });
 
       // OSS Model Runners
       // case "anythingllm_ollama":

server/utils/agents/aibitat/providers/index.js

@@ -20,6 +20,7 @@ const ApiPieProvider = require("./apipie.js");
 const XAIProvider = require("./xai.js");
 const NovitaProvider = require("./novita.js");
 const NvidiaNimProvider = require("./nvidiaNim.js");
+const PPIOProvider = require("./ppio.js");
 
 module.exports = {
   OpenAIProvider,
@@ -44,4 +45,5 @@ module.exports = {
   XAIProvider,
   NovitaProvider,
   NvidiaNimProvider,
+  PPIOProvider,
 };

server/utils/agents/aibitat/providers/ppio.js

@@ -0,0 +1,115 @@
+const OpenAI = require("openai");
+const Provider = require("./ai-provider.js");
+const InheritMultiple = require("./helpers/classes.js");
+const UnTooled = require("./helpers/untooled.js");
+
+/**
+ * The agent provider for the PPIO AI provider.
+ */
+class PPIOProvider extends InheritMultiple([Provider, UnTooled]) {
+  model;
+
+  constructor(config = {}) {
+    const { model = "qwen/qwen2.5-32b-instruct" } = config;
+    super();
+    const client = new OpenAI({
+      baseURL: "https://api.ppinfra.com/v3/openai",
+      apiKey: process.env.PPIO_API_KEY,
+      maxRetries: 3,
+      defaultHeaders: {
+        "HTTP-Referer": "https://anythingllm.com",
+        "X-API-Source": "anythingllm",
+      },
+    });
+
+    this._client = client;
+    this.model = model;
+    this.verbose = true;
+  }
+
+  get client() {
+    return this._client;
+  }
+
+  async #handleFunctionCallChat({ messages = [] }) {
+    return await this.client.chat.completions
+      .create({
+        model: this.model,
+        temperature: 0,
+        messages,
+      })
+      .then((result) => {
+        if (!Object.prototype.hasOwnProperty.call(result, "choices"))
+          throw new Error("PPIO chat: No results!");
+        if (result.choices.length === 0)
+          throw new Error("PPIO chat: No results length!");
+        return result.choices[0].message.content;
+      })
+      .catch((_) => {
+        return null;
+      });
+  }
+
+  /**
+   * Create a completion based on the received messages.
+   *
+   * @param messages A list of messages to send to the API.
+   * @param functions
+   * @returns The completion.
+   */
+  async complete(messages, functions = null) {
+    let completion;
+    if (functions.length > 0) {
+      const { toolCall, text } = await this.functionCall(
+        messages,
+        functions,
+        this.#handleFunctionCallChat.bind(this)
+      );
+
+      if (toolCall !== null) {
+        this.providerLog(`Valid tool call found - running ${toolCall.name}.`);
+        this.deduplicator.trackRun(toolCall.name, toolCall.arguments);
+        return {
+          result: null,
+          functionCall: {
+            name: toolCall.name,
+            arguments: toolCall.arguments,
+          },
+          cost: 0,
+        };
+      }
+      completion = { content: text };
+    }
+
+    if (!completion?.content) {
+      this.providerLog("Will assume chat completion without tool call inputs.");
+      const response = await this.client.chat.completions.create({
+        model: this.model,
+        messages: this.cleanMsgs(messages),
+      });
+      completion = response.choices[0].message;
+    }
+
+    // The UnTooled class inherited Deduplicator is mostly useful to prevent the agent
+    // from calling the exact same function over and over in a loop within a single chat exchange
+    // _but_ we should enable it to call previously used tools in a new chat interaction.
+    this.deduplicator.reset("runs");
+    return {
+      result: completion.content,
+      cost: 0,
+    };
+  }
+
+  /**
+   * Get the cost of the completion.
+   *
+   * @param _usage The completion to get the cost for.
+   * @returns The cost of the completion.
+   * Stubbed since PPIO has no cost basis.
+   */
+  getCost() {
+    return 0;
+  }
+}
+
+module.exports = PPIOProvider;

server/utils/agents/index.js

@@ -184,6 +184,10 @@ class AgentHandler {
             "NVIDIA NIM base path must be provided to use agents."
           );
         break;
+      case "ppio":
+        if (!process.env.PPIO_API_KEY)
+          throw new Error("PPIO API Key must be provided to use agents.");
+        break;
 
       default:
         throw new Error(
@@ -249,6 +253,8 @@ class AgentHandler {
         return process.env.NOVITA_LLM_MODEL_PREF ?? "deepseek/deepseek-r1";
       case "nvidia-nim":
         return process.env.NVIDIA_NIM_LLM_MODEL_PREF ?? null;
+      case "ppio":
+        return process.env.PPIO_MODEL_PREF ?? "qwen/qwen2.5-32b-instruct";
       default:
         return null;
     }

server/utils/helpers/customModels.js

@@ -7,6 +7,7 @@ const { ElevenLabsTTS } = require("../TextToSpeech/elevenLabs");
 const { fetchNovitaModels } = require("../AiProviders/novita");
 const { parseLMStudioBasePath } = require("../AiProviders/lmStudio");
 const { parseNvidiaNimBasePath } = require("../AiProviders/nvidiaNim");
+const { fetchPPIOModels } = require("../AiProviders/ppio");
 const { GeminiLLM } = require("../AiProviders/gemini");
 
 const SUPPORT_CUSTOM_MODELS = [
@@ -29,6 +30,7 @@ const SUPPORT_CUSTOM_MODELS = [
   "novita",
   "xai",
   "gemini",
+  "ppio",
 ];
 
 async function getCustomModels(provider = "", apiKey = null, basePath = null) {
@@ -74,6 +76,8 @@ async function getCustomModels(provider = "", apiKey = null, basePath = null) {
       return await getNvidiaNimModels(basePath);
     case "gemini":
       return await getGeminiModels(apiKey);
+    case "ppio":
+      return await getPPIOModels(apiKey);
     default:
       return { models: [], error: "Invalid provider for custom models" };
   }
@@ -571,6 +575,19 @@ async function getGeminiModels(_apiKey = null) {
   return { models, error: null };
 }
 
+async function getPPIOModels() {
+  const ppioModels = await fetchPPIOModels();
+  if (!Object.keys(ppioModels).length === 0) return { models: [], error: null };
+  const models = Object.values(ppioModels).map((model) => {
+    return {
+      id: model.id,
+      organization: model.organization,
+      name: model.name,
+    };
+  });
+  return { models, error: null };
+}
+
 module.exports = {
   getCustomModels,
 };

server/utils/helpers/index.js

@@ -197,6 +197,9 @@ function getLLMProvider({ provider = null, model = null } = {}) {
     case "nvidia-nim":
       const { NvidiaNimLLM } = require("../AiProviders/nvidiaNim");
       return new NvidiaNimLLM(embedder, model);
+    case "ppio":
+      const { PPIOLLM } = require("../AiProviders/ppio");
+      return new PPIOLLM(embedder, model);
     default:
       throw new Error(
         `ENV: No valid LLM_PROVIDER value found in environment! Using ${process.env.LLM_PROVIDER}`
@@ -338,6 +341,9 @@ function getLLMProviderClass({ provider = null } = {}) {
     case "nvidia-nim":
       const { NvidiaNimLLM } = require("../AiProviders/nvidiaNim");
       return NvidiaNimLLM;
+    case "ppio":
+      const { PPIOLLM } = require("../AiProviders/ppio");
+      return PPIOLLM;
     default:
       return null;
   }

server/utils/helpers/updateENV.js

@@ -615,6 +615,16 @@ const KEY_MAPPING = {
       },
     ],
   },
+
+  // PPIO Options
+  PPIOApiKey: {
+    envKey: "PPIO_API_KEY",
+    checks: [isNotEmpty],
+  },
+  PPIOModelPref: {
+    envKey: "PPIO_MODEL_PREF",
+    checks: [isNotEmpty],
+  },
 };
 
 function isNotEmpty(input = "") {
@@ -721,6 +731,7 @@ function supportedLLM(input = "") {
     "apipie",
     "xai",
     "nvidia-nim",
+    "ppio",
   ].includes(input);
   return validSelection ? null : `${input} is not a valid LLM provider.`;
 }