✨ feat: add Ai21Labs model provider

Dockerfile

@@ -105,6 +105,8 @@ ENV ACCESS_CODE="" \
 
 # Model Variables
 ENV \
+    # AI21
+    AI21_API_KEY="" \
     # Ai360
     AI360_API_KEY="" \
     # Anthropic

Dockerfile.database

@@ -137,6 +137,8 @@ ENV NEXT_PUBLIC_S3_DOMAIN="" \
 
 # Model Variables
 ENV \
+    # AI21
+    AI21_API_KEY="" \
     # Ai360
     AI360_API_KEY="" \
     # Anthropic

src/app/(main)/settings/llm/ProviderList/providers.tsx

@@ -1,6 +1,7 @@
 import { useMemo } from 'react';
 
 import {
+  Ai21ProviderCard,
   Ai360ProviderCard,
   AnthropicProviderCard,
   BaichuanProviderCard,
@@ -54,6 +55,7 @@ export const useProviderList = (): ProviderItem[] => {
       TogetherAIProviderCard,
       FireworksAIProviderCard,
       UpstageProviderCard,
+      Ai21ProviderCard,
       QwenProviderCard,
       SparkProviderCard,
       ZhiPuProviderCard,

src/app/api/chat/agentRuntime.ts

@@ -230,6 +230,13 @@ const getLlmOptionsFromPayload = (provider: string, payload: JWTPayload) => {
 
       const apiKey = apiKeyManager.pick(payload?.apiKey || SPARK_API_KEY);
 
+      return { apiKey };
+    }
+    case ModelProvider.Ai21: {
+      const { AI21_API_KEY } = getLLMConfig();
+
+      const apiKey = apiKeyManager.pick(payload?.apiKey || AI21_API_KEY);
+
       return { apiKey };
     }
   }

src/config/llm.ts

@@ -109,6 +109,9 @@ export const getLLMConfig = () => {
 
       ENABLED_SPARK: z.boolean(),
       SPARK_API_KEY: z.string().optional(),
+
+      ENABLED_AI21: z.boolean(),
+      AI21_API_KEY: z.string().optional(),
     },
     runtimeEnv: {
       API_KEY_SELECT_MODE: process.env.API_KEY_SELECT_MODE,
@@ -215,6 +218,9 @@ export const getLLMConfig = () => {
 
       ENABLED_SPARK: !!process.env.SPARK_API_KEY,
       SPARK_API_KEY: process.env.SPARK_API_KEY,
+
+      ENABLED_AI21: !!process.env.AI21_API_KEY,
+      AI21_API_KEY: process.env.AI21_API_KEY,
     },
   });
 };

src/config/modelProviders/ai21.ts

@@ -0,0 +1,37 @@
+import { ModelProviderCard } from '@/types/llm';
+
+// ref https://docs.ai21.com/reference/jamba-15-api-ref
+const Ai21: ModelProviderCard = {
+  chatModels: [
+    {
+      displayName: 'Jamba 1.5 Mini',
+      enabled: true,
+      functionCall: true,
+      id: 'jamba-1.5-mini',
+      pricing: {
+        input: 0.2,
+        output: 0.4,
+      },
+      tokens: 256_000,
+    },
+    {
+      displayName: 'Jamba 1.5 Large',
+      enabled: true,
+      functionCall: true,
+      id: 'jamba-1.5-large',
+      pricing: {
+        input: 2,
+        output: 8,
+      },
+      tokens: 256_000,
+    },
+  ],
+  checkModel: 'jamba-1.5-mini',
+  id: 'ai21',
+  modelList: { showModelFetcher: true },
+  modelsUrl: 'https://docs.ai21.com/reference',
+  name: 'Ai21Labs',
+  url: 'https://studio.ai21.com',
+};
+
+export default Ai21;

src/config/modelProviders/index.ts

@@ -1,5 +1,6 @@
 import { ChatModelCard, ModelProviderCard } from '@/types/llm';
 
+import Ai21Provider from './ai21';
 import Ai360Provider from './ai360';
 import AnthropicProvider from './anthropic';
 import AzureProvider from './azure';
@@ -53,6 +54,7 @@ export const LOBE_DEFAULT_MODEL_LIST: ChatModelCard[] = [
   SiliconCloudProvider.chatModels,
   UpstageProvider.chatModels,
   SparkProvider.chatModels,
+  Ai21Provider.chatModels,
 ].flat();
 
 export const DEFAULT_MODEL_PROVIDER_LIST = [
@@ -71,6 +73,7 @@ export const DEFAULT_MODEL_PROVIDER_LIST = [
   TogetherAIProvider,
   FireworksAIProvider,
   UpstageProvider,
+  Ai21Provider,
   QwenProvider,
   SparkProvider,
   ZhiPuProvider,
@@ -93,6 +96,7 @@ export const isProviderDisableBroswerRequest = (id: string) => {
   return !!provider;
 };
 
+export { default as Ai21ProviderCard } from './ai21';
 export { default as Ai360ProviderCard } from './ai360';
 export { default as AnthropicProviderCard } from './anthropic';
 export { default as AzureProviderCard } from './azure';

src/const/settings/llm.ts

@@ -1,4 +1,5 @@
 import {
+  Ai21ProviderCard,
   Ai360ProviderCard,
   AnthropicProviderCard,
   BaichuanProviderCard,
@@ -30,6 +31,10 @@ import { ModelProvider } from '@/libs/agent-runtime';
 import { UserModelProviderConfig } from '@/types/user/settings';
 
 export const DEFAULT_LLM_CONFIG: UserModelProviderConfig = {
+  ai21: {
+    enabled: false,
+    enabledModels: filterEnabledModels(Ai21ProviderCard),
+  },
   ai360: {
     enabled: false,
     enabledModels: filterEnabledModels(Ai360ProviderCard),

src/libs/agent-runtime/AgentRuntime.ts

@@ -3,6 +3,7 @@ import { ClientOptions } from 'openai';
 import type { TracePayload } from '@/const/trace';
 
 import { LobeRuntimeAI } from './BaseAI';
+import { LobeAi21AI } from './ai21';
 import { LobeAi360AI } from './ai360';
 import { LobeAnthropicAI } from './anthropic';
 import { LobeAzureOpenAI } from './azureOpenai';
@@ -116,6 +117,7 @@ class AgentRuntime {
   static async initializeWithProviderOptions(
     provider: string,
     params: Partial<{
+      ai21: Partial<ClientOptions>;
       ai360: Partial<ClientOptions>;
       anthropic: Partial<ClientOptions>;
       azure: { apiVersion?: string; apikey?: string; endpoint?: string };
@@ -282,6 +284,11 @@ class AgentRuntime {
         runtimeModel = new LobeSparkAI(params.spark);
         break;
       }
+
+      case ModelProvider.Ai21: {
+        runtimeModel = new LobeAi21AI(params.ai21);
+        break;
+      }
     }
 
     return new AgentRuntime(runtimeModel);