feat: add Avian as a new LLM provider

core/control-plane/schema.ts

@@ -19,6 +19,7 @@ const modelDescriptionSchema = z.object({
     "ovhcloud",
     "nebius",
     "siliconflow",
+    "avian",
     "tensorix",
     "scaleway",
     "watsonx",

core/llm/autodetect.ts

@@ -74,6 +74,7 @@ const PROVIDER_HANDLES_TEMPLATING: string[] = [
   "docker",
   "nous",
   "zAI",
+  "avian",
   "tensorix",
   // TODO add these, change to inverted logic so only the ones that need templating are hardcoded
   // Asksage.ts
@@ -134,6 +135,7 @@ const PROVIDER_SUPPORTS_IMAGES: string[] = [
   "ovhcloud",
   "watsonx",
   "zAI",
+  "avian",
   "tensorix",
 ];
 
@@ -253,6 +255,7 @@ const PARALLEL_PROVIDERS: string[] = [
   "vertexai",
   "function-network",
   "scaleway",
+  "avian",
   "minimax",
   "tensorix",
 ];

core/llm/llms/Avian.ts

@@ -0,0 +1,13 @@
+import { LLMOptions } from "../../index.js";
+
+import OpenAI from "./OpenAI.js";
+
+class Avian extends OpenAI {
+  static providerName = "avian";
+  static defaultOptions: Partial<LLMOptions> = {
+    apiBase: "https://api.avian.io/v1/",
+    useLegacyCompletionsEndpoint: false,
+  };
+}
+
+export default Avian;

core/llm/llms/index.ts

@@ -11,6 +11,7 @@ import { renderTemplatedString } from "../../util/handlebars/renderTemplatedStri
 import { BaseLLM } from "../index";
 import Anthropic from "./Anthropic";
 import Asksage from "./Asksage";
+import Avian from "./Avian";
 import Azure from "./Azure";
 import Bedrock from "./Bedrock";
 import BedrockImport from "./BedrockImport";
@@ -134,6 +135,7 @@ export const LLMClasses = [
   LlamaStack,
   TARS,
   zAI,
+  Avian,
 ];
 
 export async function llmFromDescription(

core/llm/toolSupport.ts

@@ -433,6 +433,15 @@ export const PROVIDER_TOOL_SUPPORT: Record<string, (model: string) => boolean> =
       const lower = model.toLowerCase();
       return !!lower.match(/^glm-[4-9]/);
     },
+    avian: (model) => {
+      const lower = model.toLowerCase();
+      return (
+        lower.includes("deepseek") ||
+        lower.includes("glm") ||
+        lower.includes("kimi") ||
+        lower.includes("minimax")
+      );
+    },
     moonshot: (model) => {
       // support moonshot models
       // https://platform.moonshot.ai/docs/pricing/chat#concepts

docs/customize/model-providers/more/avian.mdx

@@ -0,0 +1,61 @@
+---
+title: "Avian"
+description: "Configure Avian's AI models with Continue, including DeepSeek V3.2, Kimi K2.5, GLM-5, and MiniMax M2.5"
+---
+
+[Avian](https://avian.io/) provides an OpenAI-compatible API with access to leading AI models at competitive pricing.
+
+<Info>Get an API key from the [Avian dashboard](https://avian.io)</Info>
+
+## Configuration
+
+<Tabs>
+  <Tab title="YAML">
+  ```yaml title="config.yaml"
+  name: My Config
+  version: 0.0.1
+  schema: v1
+
+  models:
+    - name: DeepSeek V3.2
+      provider: avian
+      model: deepseek/deepseek-v3.2
+      apiKey: <YOUR_AVIAN_API_KEY>
+  ```
+  </Tab>
+  <Tab title="JSON (Deprecated)">
+  ```json title="config.json"
+  {
+    "models": [
+      {
+        "title": "DeepSeek V3.2",
+        "provider": "avian",
+        "model": "deepseek/deepseek-v3.2",
+        "apiKey": "<YOUR_AVIAN_API_KEY>"
+      }
+    ]
+  }
+  ```
+  </Tab>
+</Tabs>
+
+## Available Models
+
+| Model | Context Length | Input Price | Output Price |
+| ----- | ------------- | ----------- | ------------ |
+| `deepseek/deepseek-v3.2` | 164K | $0.26/M | $0.38/M |
+| `moonshotai/kimi-k2.5` | 131K | $0.45/M | $2.20/M |
+| `z-ai/glm-5` | 131K | $0.30/M | $2.55/M |
+| `minimax/minimax-m2.5` | 1M | $0.30/M | $1.10/M |
+
+## Configuration Options
+
+| Option    | Description          | Default                       |
+| --------- | -------------------- | ----------------------------- |
+| `apiKey`  | Avian API key        | Required                      |
+| `apiBase` | API base URL         | `https://api.avian.io/v1`     |
+| `model`   | Model name to use    | -                             |
+
+<Tip>
+You can set the `AVIAN_API_KEY` environment variable instead of specifying the API key directly in the configuration file.
+</Tip>

docs/docs.json

@@ -169,6 +169,7 @@
                     "group": "More Providers",
                     "pages": [
                       "customize/model-providers/more/asksage",
+                      "customize/model-providers/more/avian",
                       "customize/model-providers/more/clawrouter",
                       "customize/model-providers/more/deepseek",
                       "customize/model-providers/more/deepinfra",

extensions/vscode/config_schema.json

@@ -236,7 +236,8 @@
             "ovhcloud",
             "venice",
             "inception",
-            "tars"
+            "tars",
+            "avian"
           ],
           "markdownEnumDescriptions": [
             "### OpenAI\nUse gpt-4, gpt-3.5-turbo, or any other OpenAI model. See [here](https://openai.com/product#made-for-developers) to obtain an API key.\n\n> [Reference](https://docs.continue.dev/reference/Model%20Providers/openai)",
@@ -289,7 +290,8 @@
             "### OVHcloud AI Endpoints is a serverless inference API that provides access to a curated selection of models (e.g., Llama, Mistral, Qwen, Deepseek). It is designed with security and data privacy in mind and is compliant with GDPR. To get started, create an API key on the OVHcloud [AI Endpoints website](https://endpoints.ai.cloud.ovh.net/). For more information, including pricing, visit the OVHcloud [AI Endpoints product page](https://www.ovhcloud.com/en/public-cloud/ai-endpoints/).",
             "### Venice\n Venice.AI is a privacy-focused generative AI platform, allowing users to interact with open-source LLMs without storing any private user data.\nHosted models support the OpenAI API standard, providing seamless integration for users seeking privacy and flexibility.\nTo get started with the Venice API, either purchase a pro account, stake $VVV for daily inference allotments, or fund your account with USD.\nVisit the [API settings page](https://venice.ai/settings/api) or learn more at the [Venice API documentation](https://venice.ai/api).",
             "### Inception\n Inception Labs offer a new generation of diffusion-based LLMs.\nVisit the [API settings page](https://platform.inceptionlabs.ai/) or learn more at the [Inception docs](https://platform.inceptionlabs.ai/docs).",
-            "### TARS\nTARS is an OpenAI-compatible proxy router. To get started, obtain an API key and configure the provider in your config.json."
+            "### TARS\nTARS is an OpenAI-compatible proxy router. To get started, obtain an API key and configure the provider in your config.json.",
+            "### Avian\nAvian provides an OpenAI-compatible API with access to leading AI models including DeepSeek, Kimi, GLM, and MiniMax. To get started, obtain an API key from [avian.io](https://avian.io)."
           ],
           "type": "string"
         },
@@ -536,7 +538,8 @@
                   "kindo",
                   "scaleway",
                   "ovhcloud",
-                  "venice"
+                  "venice",
+                  "avian"
                 ]
               }
             },

gui/src/pages/AddNewModel/configs/models.ts

@@ -563,6 +563,59 @@ export const models: { [key: string]: ModelPackage } = {
     providerOptions: ["zAI"],
     isOpenSource: false,
   },
+  avianDeepseekV32: {
+    title: "DeepSeek V3.2",
+    description: "DeepSeek V3.2 with 164K context, available through Avian",
+    refUrl: "https://avian.io",
+    params: {
+      title: "DeepSeek V3.2",
+      model: "deepseek/deepseek-v3.2",
+      contextLength: 164_000,
+    },
+    icon: "avian.png",
+    providerOptions: ["avian"],
+    isOpenSource: false,
+  },
+  avianKimiK25: {
+    title: "Kimi K2.5",
+    description:
+      "Moonshot AI's Kimi K2.5 with 131K context, available through Avian",
+    refUrl: "https://avian.io",
+    params: {
+      title: "Kimi K2.5",
+      model: "moonshotai/kimi-k2.5",
+      contextLength: 131_000,
+    },
+    icon: "avian.png",
+    providerOptions: ["avian"],
+    isOpenSource: false,
+  },
+  avianGlm5: {
+    title: "GLM-5",
+    description: "Z.ai's GLM-5 with 131K context, available through Avian",
+    refUrl: "https://avian.io",
+    params: {
+      title: "GLM-5",
+      model: "z-ai/glm-5",
+      contextLength: 131_000,
+    },
+    icon: "avian.png",
+    providerOptions: ["avian"],
+    isOpenSource: false,
+  },
+  avianMinimaxM25: {
+    title: "MiniMax M2.5",
+    description: "MiniMax M2.5 with 1M context window, available through Avian",
+    refUrl: "https://avian.io",
+    params: {
+      title: "MiniMax M2.5",
+      model: "minimax/minimax-m2.5",
+      contextLength: 1_000_000,
+    },
+    icon: "avian.png",
+    providerOptions: ["avian"],
+    isOpenSource: false,
+  },
   mistralOs: {
     title: "Mistral",
     description:

gui/src/pages/AddNewModel/configs/providers.ts

@@ -266,6 +266,32 @@ export const providers: Partial<Record<string, ProviderInfo>> = {
     ],
     apiKeyUrl: "https://z.ai/manage-apikey/apikey-list",
   },
+  avian: {
+    title: "Avian",
+    provider: "avian",
+    description: "Access top AI models at low cost through Avian's API",
+    longDescription:
+      "Avian provides an OpenAI-compatible API with access to leading AI models including DeepSeek V3.2, Kimi K2.5, GLM-5, and MiniMax M2.5. Get your API key from the [Avian dashboard](https://avian.io).",
+    icon: "avian.png",
+    tags: [ModelProviderTags.RequiresApiKey],
+    packages: [
+      models.avianDeepseekV32,
+      models.avianKimiK25,
+      models.avianGlm5,
+      models.avianMinimaxM25,
+    ],
+    collectInputFor: [
+      {
+        inputType: "text",
+        key: "apiKey",
+        label: "API Key",
+        placeholder: "Enter your Avian API key",
+        required: true,
+      },
+      ...completionParamsInputsConfigs,
+    ],
+    apiKeyUrl: "https://avian.io",
+  },
   "function-network": {
     title: "Function Network",
     provider: "function-network",

packages/config-types/src/index.ts

@@ -62,6 +62,7 @@ export const modelDescriptionSchema = z.object({
     "scaleway",
     "watsonx",
     "minimax",
+    "avian",
   ]),
   model: z.string(),
   apiKey: z.string().optional(),

packages/llm-info/src/index.ts

@@ -1,4 +1,5 @@
 import { Anthropic } from "./providers/anthropic.js";
+import { Avian } from "./providers/avian.js";
 import { Azure } from "./providers/azure.js";
 import { Bedrock } from "./providers/bedrock.js";
 import { Cohere } from "./providers/cohere.js";
@@ -29,6 +30,7 @@ export const allModelProviders: ModelProvider[] = [
   MiniMax,
   xAI,
   zAI,
+  Avian,
 ];
 
 export const allLlms: LlmInfoWithProvider[] = allModelProviders.flatMap(

packages/llm-info/src/providers/avian.ts

@@ -0,0 +1,36 @@
+import { ModelProvider } from "../types.js";
+
+export const Avian: ModelProvider = {
+  models: [
+    {
+      model: "deepseek/deepseek-v3.2",
+      displayName: "DeepSeek V3.2",
+      contextLength: 164000,
+      recommendedFor: ["chat"],
+      regex: /deepseek\/deepseek-v3\.2/,
+    },
+    {
+      model: "moonshotai/kimi-k2.5",
+      displayName: "Kimi K2.5",
+      contextLength: 131000,
+      recommendedFor: ["chat"],
+      regex: /moonshotai\/kimi-k2\.5/,
+    },
+    {
+      model: "z-ai/glm-5",
+      displayName: "GLM-5",
+      contextLength: 131000,
+      recommendedFor: ["chat"],
+      regex: /z-ai\/glm-5/,
+    },
+    {
+      model: "minimax/minimax-m2.5",
+      displayName: "MiniMax M2.5",
+      contextLength: 1000000,
+      recommendedFor: ["chat"],
+      regex: /minimax\/minimax-m2\.5/,
+    },
+  ],
+  id: "avian",
+  displayName: "Avian",
+};

packages/openai-adapters/src/index.ts

@@ -133,6 +133,8 @@ export function constructLlmApi(config: LLMConfig): BaseLlmApi | undefined {
       return openAICompatible("https://api.x.ai/v1/", config);
     case "zAI":
       return openAICompatible("https://api.z.ai/api/paas/v4/", config);
+    case "avian":
+      return openAICompatible("https://api.avian.io/v1/", config);
     case "voyage":
       return openAICompatible("https://api.voyageai.com/v1/", config);
     case "mistral":

packages/openai-adapters/src/types.ts

@@ -58,6 +58,7 @@ export const OpenAIConfigSchema = BasePlusConfig.extend({
     z.literal("vllm"),
     z.literal("xAI"),
     z.literal("zAI"),
+    z.literal("avian"),
     z.literal("scaleway"),
     z.literal("tensorix"),
     z.literal("ncompass"),