add Novita AI provider

src/renderer/packages/model-setting-utils/index.ts

@@ -11,6 +11,7 @@ import GroqSettingUtil from './groq-setting-util'
 import { ModelSettingUtil } from './interface'
 import LMStudioSettingUtil from './lmstudio-setting-util'
 import MistralAISettingUtil from './mistral-ai-setting-util'
+import NovitaAISettingUtil from './novita-ai-setting-util'
 import OllamaSettingUtil from './ollama-setting-util'
 import OpenAISettingUtil from './openai-setting-util'
 import PerplexitySettingUtil from './perplexity-setting-util'
@@ -33,6 +34,7 @@ export function getModelSettingUtil(aiProvider: ModelProvider): ModelSettingUtil
     [ModelProviderEnum.VolcEngine]: VolcEngineSettingUtil,
     [ModelProviderEnum.MistralAI]: MistralAISettingUtil,
     [ModelProviderEnum.LMStudio]: LMStudioSettingUtil,
+    [ModelProviderEnum.NovitaAI]: NovitaAISettingUtil,
     [ModelProviderEnum.Perplexity]: PerplexitySettingUtil,
     [ModelProviderEnum.XAI]: XAISettingUtil,
     [ModelProviderEnum.Custom]: CustomModelSettingUtil,

src/renderer/packages/model-setting-utils/novita-ai-setting-util.ts

@@ -0,0 +1,67 @@
+import OpenAI from 'src/shared/models/openai'
+import { type ModelProvider, ModelProviderEnum, type ProviderSettings } from 'src/shared/types'
+import { createModelDependencies } from '@/adapters'
+import BaseConfig from './base-config'
+import type { ModelSettingUtil } from './interface'
+
+export default class NovitaAISettingUtil extends BaseConfig implements ModelSettingUtil {
+  public provider: ModelProvider = ModelProviderEnum.NovitaAI
+
+  async getCurrentModelDisplayName(
+    model: string,
+  ): Promise<string> {
+      return `Novita AI (${model})`
+  }
+
+  protected async listProviderModels(settings: ProviderSettings): Promise<string[]> {
+    const model = settings.models?.[0] || { modelId: 'deepseek/deepseek-v3-0324' }
+    const dependencies = await createModelDependencies()
+
+    const openai = new OpenAI(
+      {
+        apiHost: settings.apiHost!,
+        apiKey: settings.apiKey!,
+        model,
+        temperature: 0,
+        dalleStyle: 'vivid',
+        injectDefaultMetadata: false,
+        useProxy: settings.useProxy || false,
+      },
+      dependencies
+    )
+
+    try {
+      const [chatModels, embeddingModels] = await Promise.all([
+        openai.listModels(),
+        this.fetchEmbeddingModels(settings.apiHost!, settings.apiKey!),
+      ])
+
+      return [...chatModels, ...embeddingModels]
+    } catch (error) {
+      console.error('Failed to fetch Novita AI models:', error)
+      return openai.listModels()
+    }
+  }
+
+  private async fetchEmbeddingModels(apiHost: string, apiKey: string): Promise<string[]> {
+    try {
+      const response = await fetch(`${apiHost}/models?model_type=embedding`, {
+        method: 'GET',
+        headers: {
+          'Authorization': `Bearer ${apiKey}`,
+          'Content-Type': 'application/json',
+        },
+      })
+
+      if (!response.ok) {
+        throw new Error(`HTTP error! status: ${response.status}`)
+      }
+
+      const data = await response.json()
+      return data.data?.map((model: any) => model.id) || []
+    } catch (error) {
+      console.error('Failed to fetch embedding models:', error)
+      return []
+    }
+  }
+} 

src/shared/defaults.ts

@@ -695,4 +695,89 @@ export const SystemProviders: ProviderBaseInfo[] = [
       ],
     },
   },
+  {
+    id: ModelProviderEnum.NovitaAI,
+    name: 'Novita AI',
+    type: ModelProviderType.OpenAI,
+    urls: {
+      website: 'https://novita.ai',
+    },
+    defaultSettings: {
+      apiHost: 'https://api.novita.ai/v3/openai',
+      apiPath: '/chat/completions',
+      models: [
+        {
+          modelId: 'deepseek/deepseek-v3-0324',
+          contextWindow: 163840,
+          maxOutput: 163840,
+          capabilities: ['tool_use'],
+        },
+        {
+          modelId: 'deepseek/deepseek-r1-0528',
+          contextWindow: 163840,
+          maxOutput: 163840,
+          capabilities: ['tool_use'],
+        },
+        {
+          modelId: 'moonshotai/kimi-k2-instruct',
+          contextWindow: 131072,
+          maxOutput: 131072,
+          capabilities: ['tool_use'],
+        },
+        {
+          modelId: 'zai-org/glm-4.5',
+          contextWindow: 131072,
+          maxOutput: 131072,
+          capabilities: ['tool_use'],
+        },
+        {
+          modelId: 'baidu/ernie-4.5-vl-424b-a47b',
+          contextWindow: 123000,
+          maxOutput: 16000,
+          capabilities: ['tool_use', 'vision'],
+        },
+        {
+          modelId: 'qwen/qwen3-235b-a22b-thinking-2507',
+          contextWindow: 131072,
+          maxOutput: 131072,
+          capabilities: ['tool_use'],
+        }, 
+        {
+          modelId: 'qwen/qwen3-235b-a22b-instruct-2507',
+          contextWindow: 262144,
+          maxOutput: 262144,
+          capabilities: ['tool_use'],
+        },  
+        {
+          modelId: 'qwen/qwen3-30b-a3b-fp8',
+          contextWindow: 262144,
+          maxOutput: 262144,
+        }, 
+        {
+          modelId: 'qwen/qwen2.5-vl-72b-instruct',
+          contextWindow: 32768,
+          maxOutput: 32768,
+          capabilities: ['vision'],
+        },
+        {
+          modelId: 'google/gemma-3-27b-it',
+          contextWindow: 32000,
+          maxOutput: 32000,
+          capabilities: ['tool_use', 'vision'],
+        },
+        {
+          modelId: 'qwen/qwen3-embedding-8b',
+          contextWindow: 32768,
+          maxOutput: 4096,
+          type: 'embedding',
+        },
+        {
+          modelId: 'baai/bge-m3',
+          contextWindow: 8192,
+          maxOutput: 96000,
+          type: 'embedding',
+        }
+      ],
+    },
+  },
 ]

src/shared/models/index.ts

@@ -270,6 +270,23 @@ export function getModel(setting: Settings, config: Config, dependencies: ModelD
         },
         dependencies
       )
+
+    case ModelProviderEnum.NovitaAI:
+      return new CustomOpenAI(
+        {
+          apiKey: providerSetting.apiKey || '',
+          apiHost: formattedApiHost,
+          apiPath: providerSetting.apiPath || '/chat/completions',
+          model,
+          temperature: setting.temperature,
+          topP: setting.topP,
+          maxTokens: setting.maxTokens,
+          stream: setting.stream,
+          useProxy: providerSetting.useProxy,
+        },
+        dependencies
+      )
+
     default:
       if (providerBaseInfo.isCustom) {
         return new CustomOpenAI(
@@ -308,6 +325,7 @@ export const aiProviderNameHash: Record<ModelProvider, string> = {
   [ModelProviderEnum.LMStudio]: 'LM Studio API',
   [ModelProviderEnum.Perplexity]: 'Perplexity API',
   [ModelProviderEnum.XAI]: 'xAI API',
+  [ModelProviderEnum.NovitaAI]: 'Novita AI API',
   [ModelProviderEnum.Custom]: 'Custom Provider',
 }
 
@@ -383,6 +401,11 @@ export const AIModelProviderMenuOptionList = [
     label: aiProviderNameHash[ModelProviderEnum.ChatGLM6B],
     disabled: false,
   },
+  {
+    value: ModelProviderEnum.NovitaAI,
+    label: aiProviderNameHash[ModelProviderEnum.NovitaAI],
+    disabled: false,
+  },
   // {
   //     value: 'hunyuan',
   //     label: '腾讯混元',

src/shared/types.ts

@@ -255,6 +255,7 @@ export enum ModelProviderEnum {
   LMStudio = 'lm-studio',
   Perplexity = 'perplexity',
   XAI = 'xAI',
+  NovitaAI = 'novita-ai',
   Custom = 'custom',
 }
 export type ModelProvider = ModelProviderEnum | string

src/shared/utils/llm_utils.ts

@@ -116,6 +116,7 @@ export function isOpenAICompatible(providerId: string, modelId: string) {
       ModelProviderEnum.Groq,
       ModelProviderEnum.DeepSeek,
       ModelProviderEnum.LMStudio,
+      ModelProviderEnum.NovitaAI,
     ].includes(providerId as ModelProviderEnum) || providerId.startsWith('custom-provider-')
   )
 }