diff --git a/src/shared/providers/definitions/eden-ai-responses.ts b/src/shared/providers/definitions/eden-ai-responses.ts
new file mode 100644
index 0000000000..6343c06657
--- /dev/null
+++ b/src/shared/providers/definitions/eden-ai-responses.ts
@@ -0,0 +1,64 @@
+import { ModelProviderEnum, ModelProviderType } from '../../types'
+import { defineProvider } from '../registry'
+import EdenAIResponses from './models/eden-ai-responses'
+
+// Eden AI Responses API: a stateful, OpenAI-compatible alternative to chat
+// completions that stores conversation history server-side.
+// Docs: https://www.edenai.co/docs/api-reference/responses/create-response
+export const edenAIResponsesProvider = defineProvider({
+  id: ModelProviderEnum.EdenAIResponses,
+  name: 'Eden AI (Responses)',
+  type: ModelProviderType.OpenAIResponses,
+  description: 'eden-ai-responses',
+  curatedModelIds: ['@edenai', 'openai/gpt-4o', 'anthropic/claude-sonnet-4-5'],
+  urls: {
+    website: 'https://www.edenai.co/',
+    apiKey: 'https://app.edenai.run/admin/api-settings/features-preferences',
+    docs: 'https://www.edenai.co/docs/api-reference/responses/create-response',
+  },
+  defaultSettings: {
+    apiHost: 'https://api.edenai.run/v3',
+    apiPath: '/responses',
+    models: [
+      {
+        modelId: '@edenai',
+        nickname: 'Eden AI (auto-router)',
+        capabilities: ['tool_use'],
+        contextWindow: 128_000,
+      },
+      {
+        modelId: 'openai/gpt-4o',
+        nickname: 'GPT-4o',
+        capabilities: ['vision', 'tool_use'],
+        contextWindow: 128_000,
+        maxOutput: 16_384,
+      },
+      {
+        modelId: 'anthropic/claude-sonnet-4-5',
+        nickname: 'Claude Sonnet 4.5',
+        capabilities: ['vision', 'tool_use', 'reasoning'],
+        contextWindow: 200_000,
+        maxOutput: 64_000,
+      },
+    ],
+  },
+  createModel: (config) => {
+    return new EdenAIResponses(
+      {
+        apiKey: config.providerSetting.apiKey || '',
+        apiHost: config.formattedApiHost || 'https://api.edenai.run/v3',
+        apiPath: config.formattedApiPath || '/responses',
+        model: config.model,
+        temperature: config.settings.temperature,
+        topP: config.settings.topP,
+        maxOutputTokens: config.settings.maxTokens,
+        stream: config.settings.stream,
+        useProxy: config.providerSetting.useProxy,
+      },
+      config.dependencies
+    )
+  },
+  getDisplayName: (modelId, providerSettings) => {
+    return `Eden AI Responses (${providerSettings?.models?.find((m) => m.modelId === modelId)?.nickname || modelId})`
+  },
+})
diff --git a/src/shared/providers/definitions/eden-ai.ts b/src/shared/providers/definitions/eden-ai.ts
new file mode 100644
index 0000000000..0a5239870f
--- /dev/null
+++ b/src/shared/providers/definitions/eden-ai.ts
@@ -0,0 +1,89 @@
+import { ModelProviderEnum, ModelProviderType } from '../../types'
+import { defineProvider } from '../registry'
+import EdenAI from './models/eden-ai'
+
+// Eden AI exposes an OpenAI-compatible Chat Completions endpoint that routes to
+// 100+ models from many providers. Model ids use the `provider/model` format,
+// e.g. `openai/gpt-4o`. The special `@edenai` id auto-routes to the best model.
+// Docs: https://www.edenai.co/docs/api-reference/chat/chat-completions
+export const edenAIProvider = defineProvider({
+  id: ModelProviderEnum.EdenAI,
+  name: 'Eden AI',
+  type: ModelProviderType.OpenAI,
+  description: 'eden-ai',
+  curatedModelIds: [
+    '@edenai',
+    'openai/gpt-4o',
+    'openai/gpt-4o-mini',
+    'anthropic/claude-sonnet-4-5',
+    'google/gemini-2.5-pro',
+    'mistral/mistral-large-latest',
+  ],
+  urls: {
+    website: 'https://www.edenai.co/',
+    apiKey: 'https://app.edenai.run/admin/api-settings/features-preferences',
+    docs: 'https://www.edenai.co/docs/api-reference/chat/chat-completions',
+  },
+  defaultSettings: {
+    apiHost: 'https://api.edenai.run/v3',
+    models: [
+      {
+        modelId: '@edenai',
+        nickname: 'Eden AI (auto-router)',
+        capabilities: ['tool_use'],
+        contextWindow: 128_000,
+      },
+      {
+        modelId: 'openai/gpt-4o',
+        nickname: 'GPT-4o',
+        capabilities: ['vision', 'tool_use'],
+        contextWindow: 128_000,
+        maxOutput: 16_384,
+      },
+      {
+        modelId: 'openai/gpt-4o-mini',
+        nickname: 'GPT-4o mini',
+        capabilities: ['vision', 'tool_use'],
+        contextWindow: 128_000,
+        maxOutput: 16_384,
+      },
+      {
+        modelId: 'anthropic/claude-sonnet-4-5',
+        nickname: 'Claude Sonnet 4.5',
+        capabilities: ['vision', 'tool_use', 'reasoning'],
+        contextWindow: 200_000,
+        maxOutput: 64_000,
+      },
+      {
+        modelId: 'google/gemini-2.5-pro',
+        nickname: 'Gemini 2.5 Pro',
+        capabilities: ['vision', 'tool_use', 'reasoning'],
+        contextWindow: 1_048_576,
+        maxOutput: 65_536,
+      },
+      {
+        modelId: 'mistral/mistral-large-latest',
+        nickname: 'Mistral Large',
+        capabilities: ['tool_use'],
+        contextWindow: 128_000,
+        maxOutput: 8_192,
+      },
+    ],
+  },
+  createModel: (config) => {
+    return new EdenAI(
+      {
+        apiKey: config.providerSetting.apiKey || '',
+        model: config.model,
+        temperature: config.settings.temperature,
+        topP: config.settings.topP,
+        maxOutputTokens: config.settings.maxTokens,
+        stream: config.settings.stream,
+      },
+      config.dependencies
+    )
+  },
+  getDisplayName: (modelId, providerSettings) => {
+    return `Eden AI (${providerSettings?.models?.find((m) => m.modelId === modelId)?.nickname || modelId})`
+  },
+})
diff --git a/src/shared/providers/definitions/models/eden-ai-responses.ts b/src/shared/providers/definitions/models/eden-ai-responses.ts
new file mode 100644
index 0000000000..d11daa164d
--- /dev/null
+++ b/src/shared/providers/definitions/models/eden-ai-responses.ts
@@ -0,0 +1,86 @@
+import { createOpenAI } from '@ai-sdk/openai'
+import { extractReasoningMiddleware, wrapLanguageModel } from 'ai'
+import AbstractAISDKModel from '../../../models/abstract-ai-sdk'
+import { fetchRemoteModels } from '../../../models/openai-compatible'
+import type { CallChatCompletionOptions } from '../../../models/types'
+import { createFetchWithProxy } from '../../../models/utils/fetch-proxy'
+import type { ProviderModelInfo } from '../../../types'
+import type { ModelDependencies } from '../../../types/adapters'
+
+interface Options {
+  apiKey: string
+  apiHost: string
+  apiPath: string
+  model: ProviderModelInfo
+  temperature?: number
+  topP?: number
+  maxOutputTokens?: number
+  stream?: boolean
+  useProxy?: boolean
+}
+
+type FetchFunction = typeof globalThis.fetch
+
+// Eden AI's Responses endpoint is OpenAI-compatible (stateful alternative to chat
+// completions). Unlike OpenAI's `https://api.openai.com` host, Eden AI keeps the
+// version segment in the path (`https://api.edenai.run/v3/responses`), so we route
+// requests explicitly to `${apiHost}${apiPath}` instead of using the shared
+// `normalizeOpenAIResponsesHostAndPath` helper (which would append `/v1`).
+// Docs: https://www.edenai.co/docs/api-reference/responses/create-response
+export default class EdenAIResponses extends AbstractAISDKModel {
+  public name = 'Eden AI Responses'
+
+  constructor(
+    public options: Options,
+    dependencies: ModelDependencies
+  ) {
+    super(options, dependencies)
+  }
+
+  protected getCallSettings() {
+    return {
+      temperature: this.options.temperature,
+      topP: this.options.topP,
+      maxOutputTokens: this.options.maxOutputTokens,
+      stream: this.options.stream,
+    }
+  }
+
+  static isSupportTextEmbedding() {
+    return true
+  }
+
+  protected getProvider(_options: CallChatCompletionOptions, fetchFunction?: FetchFunction) {
+    return createOpenAI({
+      apiKey: this.options.apiKey,
+      baseURL: this.options.apiHost,
+      fetch: fetchFunction,
+    })
+  }
+
+  protected getChatModel(options: CallChatCompletionOptions) {
+    const { apiHost, apiPath } = this.options
+    const provider = this.getProvider(options, (_input, init) =>
+      createFetchWithProxy(this.options.useProxy, this.dependencies)(`${apiHost}${apiPath}`, init)
+    )
+    return wrapLanguageModel({
+      model: provider.responses(this.options.model.modelId),
+      middleware: extractReasoningMiddleware({ tagName: 'think' }),
+    })
+  }
+
+  public listModels() {
+    return fetchRemoteModels(
+      {
+        apiHost: this.options.apiHost,
+        apiKey: this.options.apiKey,
+        useProxy: this.options.useProxy,
+      },
+      this.dependencies
+    )
+  }
+
+  protected getImageModel() {
+    return null
+  }
+}
diff --git a/src/shared/providers/definitions/models/eden-ai.ts b/src/shared/providers/definitions/models/eden-ai.ts
new file mode 100644
index 0000000000..d3588fc030
--- /dev/null
+++ b/src/shared/providers/definitions/models/eden-ai.ts
@@ -0,0 +1,28 @@
+import OpenAICompatible, { type OpenAICompatibleSettings } from '../../../models/openai-compatible'
+import type { ModelDependencies } from '../../../types/adapters'
+
+interface Options extends OpenAICompatibleSettings {}
+
+export default class EdenAI extends OpenAICompatible {
+  public name = 'Eden AI'
+  public options: Options
+  constructor(options: Omit<Options, 'apiHost'>, dependencies: ModelDependencies) {
+    const apiHost = 'https://api.edenai.run/v3'
+    super(
+      {
+        apiKey: options.apiKey,
+        apiHost,
+        model: options.model,
+        temperature: options.temperature,
+        topP: options.topP,
+        maxOutputTokens: options.maxOutputTokens,
+        stream: options.stream,
+      },
+      dependencies
+    )
+    this.options = {
+      ...options,
+      apiHost,
+    }
+  }
+}
diff --git a/src/shared/providers/index.ts b/src/shared/providers/index.ts
index ebfa92a3ea..f6ad2de9c6 100644
--- a/src/shared/providers/index.ts
+++ b/src/shared/providers/index.ts
@@ -17,6 +17,8 @@ import './definitions/minimax'
 import './definitions/moonshot'
 import './definitions/siliconflow'
 import './definitions/openrouter'
+import './definitions/eden-ai'
+import './definitions/eden-ai-responses'
 import './definitions/ollama'
 import './definitions/lmstudio'
 import './definitions/azure'
diff --git a/src/shared/types/provider.ts b/src/shared/types/provider.ts
index 85eb01d760..367dc28a34 100644
--- a/src/shared/types/provider.ts
+++ b/src/shared/types/provider.ts
@@ -25,6 +25,8 @@ export enum ModelProviderEnum {
   Perplexity = 'perplexity',
   XAI = 'xAI',
   OpenRouter = 'openrouter',
+  EdenAI = 'eden-ai',
+  EdenAIResponses = 'eden-ai-responses',
   Bedrock = 'bedrock',
   Custom = 'custom',
 }