leigest519 · lucaszhu-hue · Jun 10, 2026 · Jun 12, 2026
diff --git a/.env.example b/.env.example
@@ -31,6 +31,13 @@
 # OPENAI_API_KEY=sk-...
 # OPENAI_BASE_URL=https://api.openai.com/v1
 # OPENAI_MODEL=gpt-4o
+#
+# Atlas Cloud is OpenAI-compatible, so you can point this same flow at it to
+# reach 300+ models (Claude, GPT, Qwen-Coder, DeepSeek, GLM, Kimi, ...) with
+# one key. See https://www.atlascloud.ai/?utm_source=github&utm_medium=link&utm_campaign=opengame
+# OPENAI_API_KEY=apikey-...
+# OPENAI_BASE_URL=https://api.atlascloud.ai/v1
+# OPENAI_MODEL=deepseek-ai/deepseek-v4-pro   # reasoning model; give it max_tokens >= 512
 
 
 # -----------------------------------------------------------------------------
@@ -42,6 +49,12 @@
 # OPENGAME_REASONING_API_KEY=sk-...
 # OPENGAME_REASONING_BASE_URL=https://api.openai.com/v1
 # OPENGAME_REASONING_MODEL=gpt-4o-mini
+#
+# Or use Atlas Cloud (one openai-compat backend for reasoning/audio/image):
+# OPENGAME_REASONING_PROVIDER=openai-compat
+# OPENGAME_REASONING_API_KEY=apikey-...
+# OPENGAME_REASONING_BASE_URL=https://api.atlascloud.ai/v1
+# OPENGAME_REASONING_MODEL=deepseek-ai/deepseek-v4-pro   # reasoning model; give it max_tokens >= 512
 
 
 # -----------------------------------------------------------------------------

diff --git a/README.md b/README.md
@@ -24,6 +24,42 @@ Jiaming Han, Tianshuo Peng, Kaixuan Fan, Manyuan Zhang, Xiangyu Yue*
   <img src="assets/opengame_teaser.png" alt="OpenGame Teaser" width="100%">
 </div>
 
+---
+
+<p align="center">
+  <a href="https://www.atlascloud.ai/?utm_source=github&utm_medium=link&utm_campaign=opengame">
+    <img src="docs/atlas-cloud-logo.png" alt="Atlas Cloud" width="200">
+  </a>
+</p>
+
+> 🎁 **[Atlas Cloud](https://www.atlascloud.ai/?utm_source=github&utm_medium=link&utm_campaign=opengame)** is a full-modal AI inference platform that gives developers a single AI API to access video generation, image generation, and LLM APIs. Instead of managing multiple vendor integrations, you connect once and get unified access to 300+ curated models across all modalities.
+>
+> Atlas Cloud exposes an **OpenAI-compatible** endpoint at `https://api.atlascloud.ai/v1`, so it plugs straight into OpenGame's existing `openai-compat` provider — point the agent runtime (`OPENAI_BASE_URL`) and any `OPENGAME_*` modality at it without writing a line of glue code. That makes it a convenient single backend for the reasoning / GDD / classifier loop, the ABC-notation audio model, and OpenAI-shaped image routes.
+>
+> Check out Atlas Cloud's new coding plan promotion for more budget-friendly API access: [https://www.atlascloud.ai/console/coding-plan](https://www.atlascloud.ai/console/coding-plan)
+
+<details>
+<summary><b>Official Atlas Cloud LLM models</b> (OpenAI-compatible, synced with <a href="https://www.atlascloud.ai/zh/models/list/llm">the model list</a>)</summary>
+
+These ids work as `OPENAI_MODEL` for the agent runtime or as `OPENGAME_REASONING_MODEL` / `OPENGAME_AUDIO_MODEL` when `*_PROVIDER=openai-compat` and `*_BASE_URL=https://api.atlascloud.ai/v1`. A solid default for the agent loop is **`deepseek-ai/deepseek-v4-pro`** — it is a reasoning model, so give it enough `max_tokens` (>= 512) or the content can come back empty with `finish_reason=length`.
+
+- **Anthropic (Claude):** `anthropic/claude-haiku-4.5-20251001`, `anthropic/claude-opus-4.8`, `anthropic/claude-sonnet-4.6`
+- **OpenAI (GPT):** `openai/gpt-5.4`, `openai/gpt-5.5`
+- **Google (Gemini):** `google/gemini-3.1-flash-lite`, `google/gemini-3.1-pro-preview`, `google/gemini-3.5-flash`
+- **Alibaba Qwen:** `qwen/qwen2.5-7b-instruct`, `Qwen/Qwen3-235B-A22B-Instruct-2507`, `qwen/qwen3-235b-a22b-thinking-2507`, `qwen/qwen3-30b-a3b`, `Qwen/Qwen3-30B-A3B-Instruct-2507`, `qwen/qwen3-30b-a3b-thinking-2507`, `qwen/qwen3-32b`, `qwen/qwen3-8b`, `Qwen/Qwen3-Coder`, `qwen/qwen3-coder-next`, `qwen/qwen3-max-2026-01-23`, `Qwen/Qwen3-Next-80B-A3B-Instruct`, `Qwen/Qwen3-Next-80B-A3B-Thinking`, `Qwen/Qwen3-VL-235B-A22B-Instruct`, `qwen/qwen3-vl-235b-a22b-thinking`, `qwen/qwen3-vl-30b-a3b-instruct`, `qwen/qwen3-vl-30b-a3b-thinking`, `qwen/qwen3-vl-8b-instruct`, `qwen/qwen3.5-122b-a10b`, `qwen/qwen3.5-27b`, `qwen/qwen3.5-35b-a3b`, `qwen/qwen3.5-397b-a17b`, `qwen/qwen3.6-35b-a3b`, `qwen/qwen3.6-plus`
+- **DeepSeek:** `deepseek-ai/deepseek-ocr`, `deepseek-ai/deepseek-r1-0528`, `deepseek-ai/DeepSeek-V3-0324`, `deepseek-ai/DeepSeek-V3.1`, `deepseek-ai/DeepSeek-V3.1-Terminus`, `deepseek-ai/deepseek-v3.2`, `deepseek-ai/DeepSeek-V3.2-Exp`, `deepseek-ai/deepseek-v4-flash`, `deepseek-ai/deepseek-v4-pro`
+- **Moonshot (Kimi):** `moonshotai/Kimi-K2-Instruct`, `moonshotai/Kimi-K2-Instruct-0905`, `moonshotai/Kimi-K2-Thinking`, `moonshotai/kimi-k2.5`, `moonshotai/kimi-k2.6`
+- **Zhipu GLM:** `zai-org/GLM-4.6`, `zai-org/glm-4.7`, `zai-org/glm-5`, `zai-org/glm-5-turbo`, `zai-org/glm-5.1`, `zai-org/glm-5v-turbo`
+- **MiniMax:** `MiniMaxAI/MiniMax-M2`, `minimaxai/minimax-m2.1`, `minimaxai/minimax-m2.5`, `minimaxai/minimax-m2.7`
+- **xAI:** `xai/grok-4.3`
+- **Kwaipilot (KAT):** `kwaipilot/kat-coder-pro-v2`
+- **Other:** `owl`
+
+For game-asset generation, Atlas Cloud also serves OpenAI-shaped text-to-image routes (e.g. `openai/gpt-image-2/text-to-image`, `qwen/qwen-image-2.0/text-to-image`) usable via `OPENGAME_IMAGE_PROVIDER=openai-compat`. The full catalog and per-model details live on the [model list page](https://www.atlascloud.ai/zh/models/list/llm).
+
+</details>
+
+---
 
 ## Abstract
 
@@ -217,6 +253,16 @@ export OPENAI_BASE_URL="https://api.openai.com/v1"     # optional
 export OPENAI_MODEL="gpt-4o"                            # optional, swap in GameCoder-27B when running it locally
 ```
 
+Because this flow is OpenAI-compatible, you can also point it at
+[Atlas Cloud](https://www.atlascloud.ai/?utm_source=github&utm_medium=link&utm_campaign=opengame)
+to reach 300+ models (Claude, GPT, Qwen-Coder, DeepSeek, GLM, Kimi, …) through one key:
+
+```bash
+export OPENAI_API_KEY="$ATLASCLOUD_API_KEY"
+export OPENAI_BASE_URL="https://api.atlascloud.ai/v1"
+export OPENAI_MODEL="deepseek-ai/deepseek-v4-pro"      # reasoning model, give it max_tokens >= 512; or Qwen/Qwen3-Coder, anthropic/claude-sonnet-4.6, ...
+```
+
 #### Asset / GDD provider keys (image, video, audio, reasoning)
 
 Beyond the main agent LLM, OpenGame's asset-generation tools talk to image,
@@ -232,10 +278,15 @@ export OPENGAME_IMAGE_API_KEY=sk-...
 ```
 
 A complete env-var reference, settings.json schema, and examples for OpenAI /
-fal.ai / OpenRouter / DashScope / Doubao live in
+Atlas Cloud / fal.ai / OpenRouter / DashScope / Doubao live in
 [`docs/users/configuration/api-keys.md`](docs/users/configuration/api-keys.md).
 A copy-paste template is at [`.env.example`](.env.example).
 
+> **Tip.** Any `openai-compat` modality (reasoning / image / audio) can target
+> [Atlas Cloud](https://www.atlascloud.ai/?utm_source=github&utm_medium=link&utm_campaign=opengame)
+> by setting `OPENGAME_<MOD>_BASE_URL=https://api.atlascloud.ai/v1` — handy if you
+> want one provider behind every modality instead of juggling several keys.
+
 OpenGame prints a one-line provider-status banner at startup so you can
 confirm which modalities are wired up before the run begins.
 

diff --git a/docs/atlas-cloud-logo.png b/docs/atlas-cloud-logo.png
diff --git a/docs/users/configuration/api-keys.md b/docs/users/configuration/api-keys.md
@@ -23,17 +23,26 @@ the interactive `/auth` command.
 
 ## Supported provider families
 
-| Provider name   | Description                                                                                              | Modalities supported           |
-| --------------- | -------------------------------------------------------------------------------------------------------- | ------------------------------ |
-| `tongyi`        | Aliyun DashScope (Wan/Qwen family).                                                                      | reasoning, image, video, audio |
-| `doubao`        | Volcengine ARK (Doubao Seedream/Seedance family).                                                        | reasoning, image, video, audio |
-| `openai-compat` | Any endpoint speaking the OpenAI REST shape (OpenAI, OpenRouter, Together, fal, Stability proxies, ...). | reasoning, image, audio        |
+| Provider name   | Description                                                                                                           | Modalities supported           |
+| --------------- | --------------------------------------------------------------------------------------------------------------------- | ------------------------------ |
+| `tongyi`        | Aliyun DashScope (Wan/Qwen family).                                                                                   | reasoning, image, video, audio |
+| `doubao`        | Volcengine ARK (Doubao Seedream/Seedance family).                                                                     | reasoning, image, video, audio |
+| `openai-compat` | Any endpoint speaking the OpenAI REST shape (OpenAI, Atlas Cloud, OpenRouter, Together, fal, Stability proxies, ...). | reasoning, image, audio        |
 
 > **Note.** `openai-compat` is intentionally **not** wired up for video,
 > because there is no stable OpenAI-shaped public video API today
 > (Sora and Veo are not part of that surface). If you select it for
 > `OPENGAME_VIDEO_PROVIDER` you'll get a clear error.
 
+> **One backend for every modality.**
+> [Atlas Cloud](https://www.atlascloud.ai/?utm_source=github&utm_medium=link&utm_campaign=opengame)
+> is an OpenAI-compatible aggregator (`https://api.atlascloud.ai/v1`) that fronts
+> 300+ models — Claude, GPT, Qwen-Coder, DeepSeek, GLM, Kimi for `reasoning`/`audio`,
+> plus OpenAI-shaped image routes for `image`. Use it through `openai-compat` when you
+> want a single key behind several modalities instead of mixing vendors. The full
+> catalog is on the [model list page](https://www.atlascloud.ai/zh/models/list/llm);
+> budget-friendly access is on the [coding plan](https://www.atlascloud.ai/console/coding-plan).
+
 ---
 
 ## Configuring with environment variables (recommended)
@@ -101,6 +110,35 @@ export OPENGAME_IMAGE_BASE_URL=https://fal.run/openai/v1
 export OPENGAME_IMAGE_MODEL=fal-ai/flux/dev
 ```
 
+### Example: Atlas Cloud as a single openai-compat backend
+
+[Atlas Cloud](https://www.atlascloud.ai/?utm_source=github&utm_medium=link&utm_campaign=opengame)
+speaks the OpenAI shape at `https://api.atlascloud.ai/v1`, so one key can drive
+the reasoning, audio, and image modalities (video still needs `tongyi`/`doubao`):
+
+```bash
+# reasoning: GDD / classifier / ABC-notation brain
+export OPENGAME_REASONING_PROVIDER=openai-compat
+export OPENGAME_REASONING_API_KEY=$ATLASCLOUD_API_KEY
+export OPENGAME_REASONING_BASE_URL=https://api.atlascloud.ai/v1
+export OPENGAME_REASONING_MODEL=deepseek-ai/deepseek-v4-pro   # reasoning model, give it max_tokens >= 512; or Qwen/Qwen3-Coder, anthropic/claude-sonnet-4.6
+
+# audio: reuse the same backend for ABC music notation
+export OPENGAME_AUDIO_PROVIDER=openai-compat
+export OPENGAME_AUDIO_API_KEY=$ATLASCLOUD_API_KEY
+export OPENGAME_AUDIO_BASE_URL=https://api.atlascloud.ai/v1
+export OPENGAME_AUDIO_MODEL=deepseek-ai/deepseek-v4-pro
+
+# image: OpenAI-shaped text-to-image route
+export OPENGAME_IMAGE_PROVIDER=openai-compat
+export OPENGAME_IMAGE_API_KEY=$ATLASCLOUD_API_KEY
+export OPENGAME_IMAGE_BASE_URL=https://api.atlascloud.ai/v1
+export OPENGAME_IMAGE_MODEL=openai/gpt-image-2/text-to-image
+```
+
+Browse the full model catalog and per-model details on the
+[model list page](https://www.atlascloud.ai/zh/models/list/llm).
+
 A copy-paste template for all variables lives at the repository root in
 [`.env.example`](../../../.env.example).