Add Skypool provider docs

docs/providers/skypool.md

@@ -0,0 +1,144 @@
+# Skypool
+
+## Overview
+
+| Property | Details |
+|-------|-------|
+| Description | Skypool Token provides OpenAI-compatible chat completions for open-weight models served through a decentralized provider network. |
+| Provider Route on LiteLLM | `skypool/` |
+| Link to Provider | [Skypool](https://skypool.xyz) |
+| Base URL | `https://a.skypool.xyz/v1` |
+| Supported Operations | `/chat/completions` |
+
+## Required Variables
+
+```python showLineNumbers title="Environment Variables"
+import os
+
+os.environ["SKYPOOL_API_KEY"] = ""  # your Skypool Consumer API key
+```
+
+You can optionally override the default base URL:
+
+```bash
+export SKYPOOL_API_BASE="https://a.skypool.xyz/v1"
+```
+
+## Usage - LiteLLM Python SDK
+
+```python showLineNumbers title="Skypool Chat Completion"
+import os
+from litellm import completion
+
+os.environ["SKYPOOL_API_KEY"] = ""  # your Skypool Consumer API key
+
+response = completion(
+    model="skypool/gemma4:26b",
+    messages=[{"role": "user", "content": "Hello, introduce Skypool in one sentence."}],
+)
+
+print(response)
+```
+
+### Streaming
+
+```python showLineNumbers title="Skypool Streaming Completion"
+import os
+from litellm import completion
+
+os.environ["SKYPOOL_API_KEY"] = ""  # your Skypool Consumer API key
+
+response = completion(
+    model="skypool/qwen3.6:35b",
+    messages=[{"role": "user", "content": "Write a short poem about distributed AI."}],
+    stream=True,
+)
+
+for chunk in response:
+    print(chunk)
+```
+
+## Usage - LiteLLM Proxy Server
+
+### 1. Save your key in the environment
+
+```bash
+export SKYPOOL_API_KEY=""
+```
+
+### 2. Add Skypool to your proxy config
+
+```yaml
+model_list:
+  - model_name: skypool-gemma4-26b
+    litellm_params:
+      model: skypool/gemma4:26b
+      api_key: os.environ/SKYPOOL_API_KEY
+```
+
+### 3. Start the proxy
+
+```bash
+litellm --config /path/to/config.yaml
+```
+
+### 4. Test it
+
+```bash
+curl --location "http://0.0.0.0:4000/chat/completions" \
+  --header "Content-Type: application/json" \
+  --data '{
+    "model": "skypool-gemma4-26b",
+    "messages": [
+      {
+        "role": "user",
+        "content": "Hello, introduce Skypool in one sentence."
+      }
+    ]
+  }'
+```
+
+## Supported Models
+
+| Model | Context Window | Max Output Tokens | Input Price | Output Price |
+|-------|---------------:|------------------:|------------:|-------------:|
+| `skypool/gemma4:26b` | 131,072 | 65,536 | 45 credits / 1M tokens | 210 credits / 1M tokens |
+| `skypool/qwen3.5:9b` | 131,072 | 65,536 | 60 credits / 1M tokens | 90 credits / 1M tokens |
+| `skypool/qwen3.6:35b` | 65,536 | 65,536 | 90 credits / 1M tokens | 480 credits / 1M tokens |
+| `skypool/qwen3.5:122b` | 131,072 | 65,536 | 150 credits / 1M tokens | 1,200 credits / 1M tokens |
+
+## Supported OpenAI Parameters
+
+Skypool supports these OpenAI-compatible chat completion parameters:
+
+| Parameter | Type | Description |
+|-----------|------|-------------|
+| `messages` | array | Required. Array of chat messages. |
+| `model` | string | Required. Skypool model ID. |
+| `stream` | boolean | Optional. Enable streaming responses. |
+| `max_tokens` | integer | Optional. Maximum tokens to generate. |
+| `temperature` | float | Optional. Sampling temperature. |
+| `top_p` | float | Optional. Nucleus sampling parameter. |
+| `stop` | string/array | Optional. Stop sequences. |
+| `seed` | integer | Optional. Deterministic sampling seed when supported by the model runtime. |
+| `tools` | array | Optional. List of available tools/functions. |
+| `tool_choice` | string/object | Optional. Control tool/function calling. |
+
+## Direct API Check
+
+You can also verify your key directly against the Skypool OpenAI-compatible endpoint:
+
+```bash
+curl --location "https://a.skypool.xyz/v1/chat/completions" \
+  --header "Authorization: Bearer $SKYPOOL_API_KEY" \
+  --header "Content-Type: application/json" \
+  --data '{
+    "model": "gemma4:26b",
+    "messages": [
+      {
+        "role": "user",
+        "content": "Hello from LiteLLM."
+      }
+    ]
+  }'
+```

sidebars.js

@@ -1029,6 +1029,7 @@ const sidebars = {
         "providers/sambanova",
         "providers/sap",
         "providers/scaleway",
+        "providers/skypool",
         "providers/stability",
         "providers/synthetic",
         "providers/snowflake",