feat(server): add fallback model and baseurl in schema (#13375)

fix AI-398

<!-- This is an auto-generated comment: release notes by coderabbit.ai
-->

## Summary by CodeRabbit

* **New Features**
* Added support for specifying fallback models for multiple AI
providers, enhancing reliability when primary models are unavailable.
* Providers can now fetch and update their list of available models
dynamically from external APIs.
* Configuration options expanded to allow custom base URLs for certain
providers.

* **Bug Fixes**
* Improved model selection logic to use fallback models if the requested
model is not available online.

* **Chores**
* Updated backend dependencies to include authentication support for
Google services.

<!-- end of auto-generated comment: release notes by coderabbit.ai -->

.docker/selfhost/schema.json

@@ -669,9 +669,16 @@
         },
         "providers.openai": {
           "type": "object",
-          "description": "The config for the openai provider.\n@default {\"apiKey\":\"\"}\n@link https://github.com/openai/openai-node",
+          "description": "The config for the openai provider.\n@default {\"apiKey\":\"\",\"baseUrl\":\"\",\"fallback\":{\"text\":\"\",\"structured\":\"\",\"image\":\"\",\"embedding\":\"\"}}\n@link https://github.com/openai/openai-node",
           "default": {
-            "apiKey": ""
+            "apiKey": "",
+            "baseUrl": "",
+            "fallback": {
+              "text": "",
+              "structured": "",
+              "image": "",
+              "embedding": ""
+            }
           }
         },
         "providers.fal": {
@@ -683,14 +690,21 @@
         },
         "providers.gemini": {
           "type": "object",
-          "description": "The config for the gemini provider.\n@default {\"apiKey\":\"\"}",
+          "description": "The config for the gemini provider.\n@default {\"apiKey\":\"\",\"baseUrl\":\"\",\"fallback\":{\"text\":\"\",\"structured\":\"\",\"image\":\"\",\"embedding\":\"\"}}",
           "default": {
-            "apiKey": ""
+            "apiKey": "",
+            "baseUrl": "",
+            "fallback": {
+              "text": "",
+              "structured": "",
+              "image": "",
+              "embedding": ""
+            }
           }
         },
         "providers.geminiVertex": {
           "type": "object",
-          "description": "The config for the google vertex provider.\n@default {}",
+          "description": "The config for the google vertex provider.\n@default {\"baseURL\":\"\",\"fallback\":{\"text\":\"\",\"structured\":\"\",\"image\":\"\",\"embedding\":\"\"}}",
           "properties": {
             "location": {
               "type": "string",
@@ -721,25 +735,39 @@
               }
             }
           },
-          "default": {}
+          "default": {
+            "baseURL": "",
+            "fallback": {
+              "text": "",
+              "structured": "",
+              "image": "",
+              "embedding": ""
+            }
+          }
         },
         "providers.perplexity": {
           "type": "object",
-          "description": "The config for the perplexity provider.\n@default {\"apiKey\":\"\"}",
+          "description": "The config for the perplexity provider.\n@default {\"apiKey\":\"\",\"fallback\":{\"text\":\"\"}}",
           "default": {
-            "apiKey": ""
+            "apiKey": "",
+            "fallback": {
+              "text": ""
+            }
           }
         },
         "providers.anthropic": {
           "type": "object",
-          "description": "The config for the anthropic provider.\n@default {\"apiKey\":\"\"}",
+          "description": "The config for the anthropic provider.\n@default {\"apiKey\":\"\",\"fallback\":{\"text\":\"\"}}",
           "default": {
-            "apiKey": ""
+            "apiKey": "",
+            "fallback": {
+              "text": ""
+            }
           }
         },
         "providers.anthropicVertex": {
           "type": "object",
-          "description": "The config for the google vertex provider.\n@default {}",
+          "description": "The config for the google vertex provider.\n@default {\"baseURL\":\"\",\"fallback\":{\"text\":\"\"}}",
           "properties": {
             "location": {
               "type": "string",
@@ -770,7 +798,12 @@
               }
             }
           },
-          "default": {}
+          "default": {
+            "baseURL": "",
+            "fallback": {
+              "text": ""
+            }
+          }
         },
         "providers.morph": {
           "type": "object",

packages/backend/server/package.json

@@ -86,6 +86,7 @@
     "express": "^5.0.1",
     "fast-xml-parser": "^5.0.0",
     "get-stream": "^9.0.1",
+    "google-auth-library": "^10.2.0",
     "graphql": "^16.9.0",
     "graphql-scalars": "^1.24.0",
     "graphql-upload": "^17.0.0",

packages/backend/server/src/plugins/copilot/config.ts

@@ -47,6 +47,13 @@ defineModuleConfig('copilot', {
     desc: 'The config for the openai provider.',
     default: {
       apiKey: '',
+      baseUrl: '',
+      fallback: {
+        text: '',
+        structured: '',
+        image: '',
+        embedding: '',
+      },
     },
     link: 'https://github.com/openai/openai-node',
   },
@@ -60,28 +67,54 @@ defineModuleConfig('copilot', {
     desc: 'The config for the gemini provider.',
     default: {
       apiKey: '',
+      baseUrl: '',
+      fallback: {
+        text: '',
+        structured: '',
+        image: '',
+        embedding: '',
+      },
     },
   },
   'providers.geminiVertex': {
     desc: 'The config for the gemini provider in Google Vertex AI.',
-    default: {},
+    default: {
+      baseURL: '',
+      fallback: {
+        text: '',
+        structured: '',
+        image: '',
+        embedding: '',
+      },
+    },
     schema: VertexSchema,
   },
   'providers.perplexity': {
     desc: 'The config for the perplexity provider.',
     default: {
       apiKey: '',
+      fallback: {
+        text: '',
+      },
     },
   },
   'providers.anthropic': {
     desc: 'The config for the anthropic provider.',
     default: {
       apiKey: '',
+      fallback: {
+        text: '',
+      },
     },
   },
   'providers.anthropicVertex': {
     desc: 'The config for the anthropic provider in Google Vertex AI.',
-    default: {},
+    default: {
+      baseURL: '',
+      fallback: {
+        text: '',
+      },
+    },
     schema: VertexSchema,
   },
   'providers.morph': {

packages/backend/server/src/plugins/copilot/providers/anthropic/official.ts

@@ -3,12 +3,23 @@ import {
   createAnthropic,
 } from '@ai-sdk/anthropic';
 
-import { CopilotProviderType, ModelInputType, ModelOutputType } from '../types';
+import {
+  CopilotChatOptions,
+  CopilotProviderType,
+  ModelConditions,
+  ModelInputType,
+  ModelOutputType,
+  PromptMessage,
+  StreamObject,
+} from '../types';
 import { AnthropicProvider } from './anthropic';
 
 export type AnthropicOfficialConfig = {
   apiKey: string;
   baseUrl?: string;
+  fallback?: {
+    text?: string;
+  };
 };
 
 export class AnthropicOfficialProvider extends AnthropicProvider<AnthropicOfficialConfig> {
@@ -67,4 +78,31 @@ export class AnthropicOfficialProvider extends AnthropicProvider<AnthropicOffici
       baseURL: this.config.baseUrl,
     });
   }
+
+  override async text(
+    cond: ModelConditions,
+    messages: PromptMessage[],
+    options: CopilotChatOptions = {}
+  ): Promise<string> {
+    const fullCond = { ...cond, fallbackModel: this.config.fallback?.text };
+    return super.text(fullCond, messages, options);
+  }
+
+  override async *streamText(
+    cond: ModelConditions,
+    messages: PromptMessage[],
+    options: CopilotChatOptions = {}
+  ): AsyncIterable<string> {
+    const fullCond = { ...cond, fallbackModel: this.config.fallback?.text };
+    yield* super.streamText(fullCond, messages, options);
+  }
+
+  override async *streamObject(
+    cond: ModelConditions,
+    messages: PromptMessage[],
+    options: CopilotChatOptions = {}
+  ): AsyncIterable<StreamObject> {
+    const fullCond = { ...cond, fallbackModel: this.config.fallback?.text };
+    yield* super.streamObject(fullCond, messages, options);
+  }
 }

packages/backend/server/src/plugins/copilot/providers/anthropic/vertex.ts

@@ -4,10 +4,23 @@ import {
   type GoogleVertexAnthropicProviderSettings,
 } from '@ai-sdk/google-vertex/anthropic';
 
-import { CopilotProviderType, ModelInputType, ModelOutputType } from '../types';
+import {
+  CopilotChatOptions,
+  CopilotProviderType,
+  ModelConditions,
+  ModelInputType,
+  ModelOutputType,
+  PromptMessage,
+  StreamObject,
+} from '../types';
+import { getGoogleAuth, VertexModelListSchema } from '../utils';
 import { AnthropicProvider } from './anthropic';
 
-export type AnthropicVertexConfig = GoogleVertexAnthropicProviderSettings;
+export type AnthropicVertexConfig = GoogleVertexAnthropicProviderSettings & {
+  fallback?: {
+    text?: string;
+  };
+};
 
 export class AnthropicVertexProvider extends AnthropicProvider<AnthropicVertexConfig> {
   override readonly type = CopilotProviderType.AnthropicVertex;
@@ -62,4 +75,54 @@ export class AnthropicVertexProvider extends AnthropicProvider<AnthropicVertexCo
     super.setup();
     this.instance = createVertexAnthropic(this.config);
   }
+
+  override async text(
+    cond: ModelConditions,
+    messages: PromptMessage[],
+    options: CopilotChatOptions = {}
+  ): Promise<string> {
+    const fullCond = { ...cond, fallbackModel: this.config.fallback?.text };
+    return super.text(fullCond, messages, options);
+  }
+
+  override async *streamText(
+    cond: ModelConditions,
+    messages: PromptMessage[],
+    options: CopilotChatOptions = {}
+  ): AsyncIterable<string> {
+    const fullCond = { ...cond, fallbackModel: this.config.fallback?.text };
+    yield* super.streamText(fullCond, messages, options);
+  }
+
+  override async *streamObject(
+    cond: ModelConditions,
+    messages: PromptMessage[],
+    options: CopilotChatOptions = {}
+  ): AsyncIterable<StreamObject> {
+    const fullCond = { ...cond, fallbackModel: this.config.fallback?.text };
+    yield* super.streamObject(fullCond, messages, options);
+  }
+
+  override async refreshOnlineModels() {
+    try {
+      const { baseUrl, headers } = await getGoogleAuth(
+        this.config,
+        'anthropic'
+      );
+      if (baseUrl && !this.onlineModelList.length) {
+        const { publisherModels } = await fetch(`${baseUrl}/models`, {
+          headers: headers(),
+        })
+          .then(r => r.json())
+          .then(r => VertexModelListSchema.parse(r));
+        this.onlineModelList = publisherModels.map(
+          model =>
+            model.name.replace('publishers/anthropic/models/', '') +
+            (model.versionId !== 'default' ? `@${model.versionId}` : '')
+        );
+      }
+    } catch (e) {
+      this.logger.error('Failed to fetch available models', e);
+    }
+  }
 }

packages/backend/server/src/plugins/copilot/providers/gemini/gemini.ts

@@ -37,11 +37,6 @@ import {
 
 export const DEFAULT_DIMENSIONS = 256;
 
-export type GeminiConfig = {
-  apiKey: string;
-  baseUrl?: string;
-};
-
 export abstract class GeminiProvider<T> extends CopilotProvider<T> {
   private readonly MAX_STEPS = 20;
 

packages/backend/server/src/plugins/copilot/providers/gemini/generative.ts

@@ -2,15 +2,35 @@ import {
   createGoogleGenerativeAI,
   type GoogleGenerativeAIProvider,
 } from '@ai-sdk/google';
+import z from 'zod';
 
-import { CopilotProviderType, ModelInputType, ModelOutputType } from '../types';
+import {
+  CopilotChatOptions,
+  CopilotEmbeddingOptions,
+  CopilotProviderType,
+  ModelConditions,
+  ModelInputType,
+  ModelOutputType,
+  PromptMessage,
+  StreamObject,
+} from '../types';
 import { GeminiProvider } from './gemini';
 
 export type GeminiGenerativeConfig = {
   apiKey: string;
   baseUrl?: string;
+  fallback?: {
+    text?: string;
+    structured?: string;
+    image?: string;
+    embedding?: string;
+  };
 };
 
+const ModelListSchema = z.object({
+  models: z.array(z.object({ name: z.string() })),
+});
+
 export class GeminiGenerativeProvider extends GeminiProvider<GeminiGenerativeConfig> {
   override readonly type = CopilotProviderType.Gemini;
 
@@ -71,27 +91,16 @@ export class GeminiGenerativeProvider extends GeminiProvider<GeminiGenerativeCon
       ],
     },
     {
-      name: 'Text Embedding 005',
+      name: 'Gemini Embedding',
-      id: 'text-embedding-005',
+      id: 'gemini-embedding-001',
       capabilities: [
         {
           input: [ModelInputType.Text],
           output: [ModelOutputType.Embedding],
+          defaultForOutputType: true,
         },
       ],
     },
-    // not exists yet
-    // {
-    //   name: 'Gemini Embedding',
-    //   id: 'gemini-embedding-001',
-    //   capabilities: [
-    //     {
-    //       input: [ModelInputType.Text],
-    //       output: [ModelOutputType.Embedding],
-    //       defaultForOutputType: true,
-    //     },
-    //   ],
-    // },
   ];
 
   protected instance!: GoogleGenerativeAIProvider;
@@ -107,4 +116,77 @@ export class GeminiGenerativeProvider extends GeminiProvider<GeminiGenerativeCon
       baseURL: this.config.baseUrl,
     });
   }
+
+  override async text(
+    cond: ModelConditions,
+    messages: PromptMessage[],
+    options: CopilotChatOptions = {}
+  ): Promise<string> {
+    const fullCond = { ...cond, fallbackModel: this.config.fallback?.text };
+    return super.text(fullCond, messages, options);
+  }
+
+  override async structure(
+    cond: ModelConditions,
+    messages: PromptMessage[],
+    options?: CopilotChatOptions
+  ): Promise<string> {
+    const fullCond = {
+      ...cond,
+      fallbackModel: this.config.fallback?.structured,
+    };
+    return super.structure(fullCond, messages, options);
+  }
+
+  override async *streamText(
+    cond: ModelConditions,
+    messages: PromptMessage[],
+    options: CopilotChatOptions = {}
+  ): AsyncIterable<string> {
+    const fullCond = { ...cond, fallbackModel: this.config.fallback?.text };
+    yield* super.streamText(fullCond, messages, options);
+  }
+
+  override async *streamObject(
+    cond: ModelConditions,
+    messages: PromptMessage[],
+    options: CopilotChatOptions = {}
+  ): AsyncIterable<StreamObject> {
+    const fullCond = { ...cond, fallbackModel: this.config.fallback?.text };
+    yield* super.streamObject(fullCond, messages, options);
+  }
+
+  override async embedding(
+    cond: ModelConditions,
+    messages: string | string[],
+    options?: CopilotEmbeddingOptions
+  ): Promise<number[][]> {
+    const fullCond = {
+      ...cond,
+      fallbackModel: this.config.fallback?.embedding,
+    };
+    return super.embedding(fullCond, messages, options);
+  }
+
+  override async refreshOnlineModels() {
+    try {
+      const baseUrl =
+        this.config.baseUrl ||
+        'https://generativelanguage.googleapis.com/v1beta';
+      if (baseUrl && !this.onlineModelList.length) {
+        const { models } = await fetch(
+          `${baseUrl}/models?key=${this.config.apiKey}`
+        )
+          .then(r => r.json())
+          .then(
+            r => (console.log(JSON.stringify(r)), ModelListSchema.parse(r))
+          );
+        this.onlineModelList = models.map(model =>
+          model.name.replace('models/', '')
+        );
+      }
+    } catch (e) {
+      this.logger.error('Failed to fetch available models', e);
+    }
+  }
 }

packages/backend/server/src/plugins/copilot/providers/gemini/vertex.ts

@@ -4,10 +4,27 @@ import {
   type GoogleVertexProviderSettings,
 } from '@ai-sdk/google-vertex';
 
-import { CopilotProviderType, ModelInputType, ModelOutputType } from '../types';
+import {
+  CopilotChatOptions,
+  CopilotEmbeddingOptions,
+  CopilotProviderType,
+  ModelConditions,
+  ModelInputType,
+  ModelOutputType,
+  PromptMessage,
+  StreamObject,
+} from '../types';
+import { getGoogleAuth, VertexModelListSchema } from '../utils';
 import { GeminiProvider } from './gemini';
 
-export type GeminiVertexConfig = GoogleVertexProviderSettings;
+export type GeminiVertexConfig = GoogleVertexProviderSettings & {
+  fallback?: {
+    text?: string;
+    structured?: string;
+    image?: string;
+    embedding?: string;
+  };
+};
 
 export class GeminiVertexProvider extends GeminiProvider<GeminiVertexConfig> {
   override readonly type = CopilotProviderType.GeminiVertex;
@@ -72,4 +89,73 @@ export class GeminiVertexProvider extends GeminiProvider<GeminiVertexConfig> {
     super.setup();
     this.instance = createVertex(this.config);
   }
+
+  override async text(
+    cond: ModelConditions,
+    messages: PromptMessage[],
+    options: CopilotChatOptions = {}
+  ): Promise<string> {
+    const fullCond = { ...cond, fallbackModel: this.config.fallback?.text };
+    return super.text(fullCond, messages, options);
+  }
+
+  override async structure(
+    cond: ModelConditions,
+    messages: PromptMessage[],
+    options?: CopilotChatOptions
+  ): Promise<string> {
+    const fullCond = {
+      ...cond,
+      fallbackModel: this.config.fallback?.structured,
+    };
+    return super.structure(fullCond, messages, options);
+  }
+
+  override async *streamText(
+    cond: ModelConditions,
+    messages: PromptMessage[],
+    options: CopilotChatOptions = {}
+  ): AsyncIterable<string> {
+    const fullCond = { ...cond, fallbackModel: this.config.fallback?.text };
+    yield* super.streamText(fullCond, messages, options);
+  }
+
+  override async *streamObject(
+    cond: ModelConditions,
+    messages: PromptMessage[],
+    options: CopilotChatOptions = {}
+  ): AsyncIterable<StreamObject> {
+    const fullCond = { ...cond, fallbackModel: this.config.fallback?.text };
+    yield* super.streamObject(fullCond, messages, options);
+  }
+
+  override async embedding(
+    cond: ModelConditions,
+    messages: string | string[],
+    options?: CopilotEmbeddingOptions
+  ): Promise<number[][]> {
+    const fullCond = {
+      ...cond,
+      fallbackModel: this.config.fallback?.embedding,
+    };
+    return super.embedding(fullCond, messages, options);
+  }
+
+  override async refreshOnlineModels() {
+    try {
+      const { baseUrl, headers } = await getGoogleAuth(this.config, 'google');
+      if (baseUrl && !this.onlineModelList.length) {
+        const { publisherModels } = await fetch(`${baseUrl}/models`, {
+          headers: headers(),
+        })
+          .then(r => r.json())
+          .then(r => VertexModelListSchema.parse(r));
+        this.onlineModelList = publisherModels.map(model =>
+          model.name.replace('publishers/google/models/', '')
+        );
+      }
+    } catch (e) {
+      this.logger.error('Failed to fetch available models', e);
+    }
+  }
 }

packages/backend/server/src/plugins/copilot/providers/openai.ts

@@ -46,8 +46,18 @@ export const DEFAULT_DIMENSIONS = 256;
 export type OpenAIConfig = {
   apiKey: string;
   baseUrl?: string;
+  fallback?: {
+    text?: string;
+    structured?: string;
+    image?: string;
+    embedding?: string;
+  };
 };
 
+const ModelListSchema = z.object({
+  data: z.array(z.object({ id: z.string() })),
+});
+
 const ImageResponseSchema = z.union([
   z.object({
     data: z.array(z.object({ b64_json: z.string() })),
@@ -271,6 +281,25 @@ export class OpenAIProvider extends CopilotProvider<OpenAIConfig> {
     }
   }
 
+  override async refreshOnlineModels() {
+    try {
+      const baseUrl = this.config.baseUrl || 'https://api.openai.com/v1';
+      if (baseUrl && !this.onlineModelList.length) {
+        const { data } = await fetch(`${baseUrl}/models`, {
+          headers: {
+            Authorization: `Bearer ${this.config.apiKey}`,
+            'Content-Type': 'application/json',
+          },
+        })
+          .then(r => r.json())
+          .then(r => ModelListSchema.parse(r));
+        this.onlineModelList = data.map(model => model.id);
+      }
+    } catch (e) {
+      this.logger.error('Failed to fetch available models', e);
+    }
+  }
+
   override getProviderSpecificTools(
     toolName: CopilotChatTools,
     model: string
@@ -291,6 +320,7 @@ export class OpenAIProvider extends CopilotProvider<OpenAIConfig> {
     const fullCond = {
       ...cond,
       outputType: ModelOutputType.Text,
+      fallbackModel: this.config.fallback?.text,
     };
     await this.checkParams({ messages, cond: fullCond, options });
     const model = this.selectModel(fullCond);
@@ -331,6 +361,7 @@ export class OpenAIProvider extends CopilotProvider<OpenAIConfig> {
     const fullCond = {
       ...cond,
       outputType: ModelOutputType.Text,
+      fallbackModel: this.config.fallback?.text,
     };
     await this.checkParams({ messages, cond: fullCond, options });
     const model = this.selectModel(fullCond);
@@ -376,7 +407,11 @@ export class OpenAIProvider extends CopilotProvider<OpenAIConfig> {
     messages: PromptMessage[],
     options: CopilotChatOptions = {}
   ): AsyncIterable<StreamObject> {
-    const fullCond = { ...cond, outputType: ModelOutputType.Object };
+    const fullCond = {
+      ...cond,
+      outputType: ModelOutputType.Object,
+      fallbackModel: this.config.fallback?.text,
+    };
     await this.checkParams({ cond: fullCond, messages, options });
     const model = this.selectModel(fullCond);
 
@@ -409,7 +444,11 @@ export class OpenAIProvider extends CopilotProvider<OpenAIConfig> {
     messages: PromptMessage[],
     options: CopilotStructuredOptions = {}
   ): Promise<string> {
-    const fullCond = { ...cond, outputType: ModelOutputType.Structured };
+    const fullCond = {
+      ...cond,
+      outputType: ModelOutputType.Structured,
+      fallbackModel: this.config.fallback?.structured,
+    };
     await this.checkParams({ messages, cond: fullCond, options });
     const model = this.selectModel(fullCond);
 
@@ -449,7 +488,11 @@ export class OpenAIProvider extends CopilotProvider<OpenAIConfig> {
     chunkMessages: PromptMessage[][],
     options: CopilotChatOptions = {}
   ): Promise<number[]> {
-    const fullCond = { ...cond, outputType: ModelOutputType.Text };
+    const fullCond = {
+      ...cond,
+      outputType: ModelOutputType.Text,
+      fallbackModel: this.config.fallback?.text,
+    };
     await this.checkParams({ messages: [], cond: fullCond, options });
     const model = this.selectModel(fullCond);
     // get the log probability of "yes"/"no"
@@ -594,7 +637,11 @@ export class OpenAIProvider extends CopilotProvider<OpenAIConfig> {
     messages: PromptMessage[],
     options: CopilotImageOptions = {}
   ) {
-    const fullCond = { ...cond, outputType: ModelOutputType.Image };
+    const fullCond = {
+      ...cond,
+      outputType: ModelOutputType.Image,
+      fallbackModel: this.config.fallback?.image,
+    };
     await this.checkParams({ messages, cond: fullCond, options });
     const model = this.selectModel(fullCond);
 
@@ -644,7 +691,11 @@ export class OpenAIProvider extends CopilotProvider<OpenAIConfig> {
     options: CopilotEmbeddingOptions = { dimensions: DEFAULT_DIMENSIONS }
   ): Promise<number[][]> {
     messages = Array.isArray(messages) ? messages : [messages];
-    const fullCond = { ...cond, outputType: ModelOutputType.Embedding };
+    const fullCond = {
+      ...cond,
+      outputType: ModelOutputType.Embedding,
+      fallbackModel: this.config.fallback?.embedding,
+    };
     await this.checkParams({ embeddings: messages, cond: fullCond, options });
     const model = this.selectModel(fullCond);
 

packages/backend/server/src/plugins/copilot/providers/perplexity.ts

@@ -20,6 +20,9 @@ import { chatToGPTMessage, CitationParser } from './utils';
 export type PerplexityConfig = {
   apiKey: string;
   endpoint?: string;
+  fallback?: {
+    text?: string;
+  };
 };
 
 const PerplexityErrorSchema = z.union([
@@ -109,7 +112,11 @@ export class PerplexityProvider extends CopilotProvider<PerplexityConfig> {
     messages: PromptMessage[],
     options: CopilotChatOptions = {}
   ): Promise<string> {
-    const fullCond = { ...cond, outputType: ModelOutputType.Text };
+    const fullCond = {
+      ...cond,
+      outputType: ModelOutputType.Text,
+      fallbackModel: this.config.fallback?.text,
+    };
     await this.checkParams({ cond: fullCond, messages, options });
     const model = this.selectModel(fullCond);
 
@@ -149,7 +156,11 @@ export class PerplexityProvider extends CopilotProvider<PerplexityConfig> {
     messages: PromptMessage[],
     options: CopilotChatOptions = {}
   ): AsyncIterable<string> {
-    const fullCond = { ...cond, outputType: ModelOutputType.Text };
+    const fullCond = {
+      ...cond,
+      outputType: ModelOutputType.Text,
+      fallbackModel: this.config.fallback?.text,
+    };
     await this.checkParams({ cond: fullCond, messages, options });
     const model = this.selectModel(fullCond);
 

packages/backend/server/src/plugins/copilot/providers/provider.ts

@@ -53,6 +53,7 @@ import {
 @Injectable()
 export abstract class CopilotProvider<C = any> {
   protected readonly logger = new Logger(this.constructor.name);
+  protected onlineModelList: string[] = [];
   abstract readonly type: CopilotProviderType;
   abstract readonly models: CopilotProviderModel[];
   abstract configured(): boolean;
@@ -80,11 +81,18 @@ export abstract class CopilotProvider<C = any> {
   protected setup() {
     if (this.configured()) {
       this.factory.register(this);
+      if (env.selfhosted) {
+        this.refreshOnlineModels().catch(e =>
+          this.logger.error('Failed to refresh online models', e)
+        );
+      }
     } else {
       this.factory.unregister(this);
     }
   }
 
+  async refreshOnlineModels() {}
+
   private findValidModel(
     cond: ModelFullConditions
   ): CopilotProviderModel | undefined {
@@ -95,9 +103,26 @@ export abstract class CopilotProvider<C = any> {
         inputTypes.every(type => cap.input.includes(type)));
 
     if (modelId) {
-      return this.models.find(
+      const hasOnlineModel = this.onlineModelList.includes(modelId);
+      const hasFallbackModel = cond.fallbackModel
+        ? this.onlineModelList.includes(cond.fallbackModel)
+        : undefined;
+
+      const model = this.models.find(
         m => m.id === modelId && m.capabilities.some(matcher)
       );
+
+      if (model) {
+        // return fallback model if current model is not alive
+        if (!hasOnlineModel && hasFallbackModel) {
+          // oxlint-disable-next-line typescript-eslint(no-non-null-assertion)
+          return { id: cond.fallbackModel!, capabilities: [] };
+        }
+        return model;
+      }
+      // allow online model without capabilities check
+      if (hasOnlineModel) return { id: modelId, capabilities: [] };
+      return undefined;
     }
     if (!outputType) return undefined;
 

packages/backend/server/src/plugins/copilot/providers/types.ts

@@ -237,6 +237,7 @@ export interface ModelCapability {
 
 export interface CopilotProviderModel {
   id: string;
+  name?: string;
   capabilities: ModelCapability[];
 }
 
@@ -247,4 +248,5 @@ export type ModelConditions = {
 
 export type ModelFullConditions = ModelConditions & {
   outputType?: ModelOutputType;
+  fallbackModel?: string;
 };

packages/backend/server/src/plugins/copilot/providers/utils.ts

@@ -1,3 +1,5 @@
+import { GoogleVertexProviderSettings } from '@ai-sdk/google-vertex';
+import { GoogleVertexAnthropicProviderSettings } from '@ai-sdk/google-vertex/anthropic';
 import { Logger } from '@nestjs/common';
 import {
   CoreAssistantMessage,
@@ -7,7 +9,8 @@ import {
   TextPart,
   TextStreamPart,
 } from 'ai';
-import { ZodType } from 'zod';
+import { GoogleAuth, GoogleAuthOptions } from 'google-auth-library';
+import z, { ZodType } from 'zod';
 
 import { CustomAITools } from '../tools';
 import { PromptMessage, StreamObject } from './types';
@@ -655,3 +658,54 @@ export class StreamObjectParser {
     }, '');
   }
 }
+
+export const VertexModelListSchema = z.object({
+  publisherModels: z.array(
+    z.object({
+      name: z.string(),
+      versionId: z.string(),
+    })
+  ),
+});
+
+export async function getGoogleAuth(
+  options: GoogleVertexAnthropicProviderSettings | GoogleVertexProviderSettings,
+  publisher: 'anthropic' | 'google'
+) {
+  function getBaseUrl() {
+    const { baseURL, location } = options;
+    if (baseURL?.trim()) {
+      try {
+        const url = new URL(baseURL);
+        if (url.pathname.endsWith('/')) {
+          url.pathname = url.pathname.slice(0, -1);
+        }
+        return url.toString();
+      } catch {}
+    } else if (location) {
+      return `https://${location}-aiplatform.googleapis.com/v1beta1/publishers/${publisher}`;
+    }
+    return undefined;
+  }
+
+  async function generateAuthToken() {
+    if (!options.googleAuthOptions) {
+      return undefined;
+    }
+    const auth = new GoogleAuth({
+      scopes: ['https://www.googleapis.com/auth/cloud-platform'],
+      ...(options.googleAuthOptions as GoogleAuthOptions),
+    });
+    const client = await auth.getClient();
+    const token = await client.getAccessToken();
+    return token.token;
+  }
+
+  const token = await generateAuthToken();
+
+  return {
+    baseUrl: getBaseUrl(),
+    headers: () => ({ Authorization: `Bearer ${token}` }),
+    fetch: options.fetch,
+  };
+}

yarn.lock

@@ -998,6 +998,7 @@ __metadata:
     express: "npm:^5.0.1"
     fast-xml-parser: "npm:^5.0.0"
     get-stream: "npm:^9.0.1"
+    google-auth-library: "npm:^10.2.0"
     graphql: "npm:^16.9.0"
     graphql-scalars: "npm:^1.24.0"
     graphql-upload: "npm:^17.0.0"
@@ -22147,14 +22148,14 @@ __metadata:
   languageName: node
   linkType: hard
 
-"gaxios@npm:^7.0.0-rc.1, gaxios@npm:^7.0.0-rc.4":
+"gaxios@npm:^7.0.0, gaxios@npm:^7.0.0-rc.4":
-  version: 7.0.0-rc.6
+  version: 7.1.1
-  resolution: "gaxios@npm:7.0.0-rc.6"
+  resolution: "gaxios@npm:7.1.1"
   dependencies:
     extend: "npm:^3.0.2"
     https-proxy-agent: "npm:^7.0.1"
     node-fetch: "npm:^3.3.2"
-  checksum: 10/60c688d4c65062c97bf0f33f959713df106e207065586bf5deb546ef5d02cddcba46d138c0b7eb8712950ca880fa28d3665936b19156224f8c478d9c4f817aea
+  checksum: 10/9e5fa8b458c318a95d4dff0f6ac187a1b8933fb1de5b376b7098b27dfc5bf6025b62c87ed20bdae0496ae73a279834bc6b974c28849a674deed0089f2ba57b98
   languageName: node
   linkType: hard
 
@@ -22169,14 +22170,14 @@ __metadata:
   languageName: node
   linkType: hard
 
-"gcp-metadata@npm:^7.0.0-rc.1":
+"gcp-metadata@npm:^7.0.0":
-  version: 7.0.0-rc.1
+  version: 7.0.1
-  resolution: "gcp-metadata@npm:7.0.0-rc.1"
+  resolution: "gcp-metadata@npm:7.0.1"
   dependencies:
-    gaxios: "npm:^7.0.0-rc.1"
+    gaxios: "npm:^7.0.0"
     google-logging-utils: "npm:^1.0.0"
     json-bigint: "npm:^1.0.0"
-  checksum: 10/2c58401c7945c41144bc6a44a066c050d36c34ee10e04e85ffde488afec6a3f67ebe29e697d0328aa20181b38a36aae9165896c0201387ea8cff031cdb790ab9
+  checksum: 10/c82f20a4ce22278998fe033e668a66bff04d2b3e95e19f968adeac829e12274e07b453fcfcf34573a6d702b3570c5556cba6eb6b59d1c03757c866e3271972c1
   languageName: node
   linkType: hard
 
@@ -22540,18 +22541,18 @@ __metadata:
   languageName: node
   linkType: hard
 
-"google-auth-library@npm:^10.0.0-rc.1":
+"google-auth-library@npm:^10.0.0-rc.1, google-auth-library@npm:^10.2.0":
-  version: 10.0.0-rc.3
+  version: 10.2.0
-  resolution: "google-auth-library@npm:10.0.0-rc.3"
+  resolution: "google-auth-library@npm:10.2.0"
   dependencies:
     base64-js: "npm:^1.3.0"
     ecdsa-sig-formatter: "npm:^1.0.11"
-    gaxios: "npm:^7.0.0-rc.4"
+    gaxios: "npm:^7.0.0"
-    gcp-metadata: "npm:^7.0.0-rc.1"
+    gcp-metadata: "npm:^7.0.0"
     google-logging-utils: "npm:^1.0.0"
-    gtoken: "npm:^8.0.0-rc.1"
+    gtoken: "npm:^8.0.0"
     jws: "npm:^4.0.0"
-  checksum: 10/d76f470ddba1d5ec84cb72b03af388722db2987a95f64c7d91169f3f67608528a327fd132bb00d74be10b08e6446906b16983cf0aa9f6fa10ac7a6d3aacff0d7
+  checksum: 10/dfa6ad7240da3915b7e15d1f39cd6906e6714502b09957be07d2429665a44c6b46d8ca7a077cec7fcb83f82ed2c3cb5a9b7c1ba79ebb3e8920eba286a24bdd63
   languageName: node
   linkType: hard
 
@@ -22748,13 +22749,13 @@ __metadata:
   languageName: node
   linkType: hard
 
-"gtoken@npm:^8.0.0-rc.1":
+"gtoken@npm:^8.0.0":
-  version: 8.0.0-rc.1
+  version: 8.0.0
-  resolution: "gtoken@npm:8.0.0-rc.1"
+  resolution: "gtoken@npm:8.0.0"
   dependencies:
-    gaxios: "npm:^7.0.0-rc.1"
+    gaxios: "npm:^7.0.0"
     jws: "npm:^4.0.0"
-  checksum: 10/d2481344df8d9f62ec3ae7fe97b562c93dc294c7d1e7d8e1603162fe2726cfb5993f2b1a4e04388ee0f49c5fd02c1b0799dc58cebbf8af10489af8de80a72902
+  checksum: 10/b921430395dcd06ee63c3fc5a5e339ca4d6dcb38b6d618beb0f260bae1088d53d130f86029a9d578f1601c64685f49a65dba57bbd617c4b14039180b67b6c5ce
   languageName: node
   linkType: hard