memodb-io · octo-patch · Jul 13, 2026
@@ -11,18 +11,20 @@ Configure your acontext core services using these essential environment variable
 
 <ParamField path="LLM_API_KEY" type="string" required>
 API key for your LLM provider (OpenAI or Anthropic). This is the primary authentication credential for AI model access.
+Use your MiniMax API key with either compatible protocol.
 </ParamField>
 
 <ParamField path="LLM_BASE_URL" type="string" default="null">
 Custom base URL for LLM API endpoints. Leave unset to use the provider's default endpoint.
 </ParamField>
 
 <ParamField path="LLM_SDK" type="string" default="openai">
-LLM provider to use. Supported values: `openai`, `anthropic`
+LLM protocol adapter to use. Supported values: `openai`, `anthropic`, `minimax`. The `minimax` value uses the OpenAI-compatible adapter.
 </ParamField>
 
 <ParamField path="LLM_SIMPLE_MODEL" type="string" default="gpt-4.1">
 Default model identifier for LLM operations. Examples: `gpt-4`, `gpt-3.5-turbo`, `claude-3-sonnet`
+For MiniMax, use `MiniMax-M3` or `MiniMax-M2.7`.
 </ParamField>
 
 <ParamField path="LLM_RESPONSE_TIMEOUT" type="float" default="60">
@@ -78,6 +80,38 @@ BLOCK_EMBEDDING_DIM=1536
 BLOCK_EMBEDDING_SEARCH_COSINE_DISTANCE_THRESHOLD=0.8
 ```
 
+```bash title="MiniMax Global - OpenAI"
+LLM_API_KEY=your-minimax-api-key
+LLM_BASE_URL=https://api.minimax.io/v1
+LLM_SDK=minimax
+LLM_SIMPLE_MODEL=MiniMax-M3
+LLM_RESPONSE_TIMEOUT=60
+```
+
+```bash title="MiniMax Global - Anthropic"
+LLM_API_KEY=your-minimax-api-key
+LLM_BASE_URL=https://api.minimax.io/anthropic
+LLM_SDK=anthropic
+LLM_SIMPLE_MODEL=MiniMax-M3
+LLM_RESPONSE_TIMEOUT=60
+```
+
+```bash title="MiniMax China - OpenAI"
+LLM_API_KEY=your-minimax-api-key
+LLM_BASE_URL=https://api.minimaxi.com/v1
+LLM_SDK=minimax
+LLM_SIMPLE_MODEL=MiniMax-M2.7
+LLM_RESPONSE_TIMEOUT=60
+```
+
+```bash title="MiniMax China - Anthropic"
+LLM_API_KEY=your-minimax-api-key
+LLM_BASE_URL=https://api.minimaxi.com/anthropic
+LLM_SDK=anthropic
+LLM_SIMPLE_MODEL=MiniMax-M2.7
+LLM_RESPONSE_TIMEOUT=60
+```
+
 ```bash title="Anthropic Setup"
 # Using Anthropic Claude
 LLM_API_KEY=your-anthropic-api-key
@@ -117,6 +151,25 @@ BLOCK_EMBEDDING_DIM=1024
 ```
 </CodeGroup>
 
+### MiniMax Models
+
+| Model | Context window | Input modalities | Thinking |
+| --- | ---: | --- | --- |
+| `MiniMax-M3` | 1,000,000 tokens | Text, image, video | Adaptive or disabled |
+| `MiniMax-M2.7` | 204,800 tokens | Text | Always on |
+
+Global pay-as-you-go prices are in USD per million tokens.
+
+| Model | Service tier | Input length | Input | Output | Cache read | Cache write |
+| --- | --- | --- | ---: | ---: | ---: | ---: |
+| `MiniMax-M3` | Standard | Up to 512,000 tokens | $0.30 | $1.20 | $0.06 | Not available |
+| `MiniMax-M3` | Standard | More than 512,000 tokens | $0.60 | $2.40 | $0.12 | Not available |
+| `MiniMax-M3` | Priority | Up to 512,000 tokens | $0.45 | $1.80 | $0.09 | Not available |
+| `MiniMax-M3` | Priority | More than 512,000 tokens | $0.90 | $3.60 | $0.18 | Not available |
+| `MiniMax-M2.7` | Default | Any supported length | $0.30 | $1.20 | $0.06 | $0.375 |
+
+See the [MiniMax model reference](https://platform.minimax.io/docs/api-reference/api-overview) and [current pricing](https://platform.minimax.io/docs/guides/pricing-paygo) for updates.
+
 ## Appendix
 
 <AccordionGroup>

@@ -756,9 +756,10 @@ func runServerUp(cmd *cobra.Command, args []string) error {
 		if _, err := os.Stat(envFile); os.IsNotExist(err) {
 			envConfig := &docker.EnvConfig{
 				LLMConfig: &docker.LLMConfig{
-					APIKey:  "your-api-key",
-					BaseURL: "https://api.openai.com/v1",
-					SDK:     "openai",
+					APIKey:      "your-api-key",
+					BaseURL:     "https://api.openai.com/v1",
+					SimpleModel: "gpt-4.1",
+					SDK:         "openai",
 				},
 				RootAPIBearerToken: "your-root-api-bearer-token",
 				CoreConfigYAMLFile: "./config.yaml",

@@ -12,6 +12,7 @@ func GenerateEnvFile(filePath string, config *EnvConfig) error {
 # LLM Configuration
 LLM_API_KEY={{.LLMAPIKey}}
 LLM_BASE_URL={{.LLMBaseURL}}
+LLM_SIMPLE_MODEL={{.LLMSimpleModel}}
 LLM_SDK={{.LLMSDK}}
 
 # API Bearer Token (for root API access)
@@ -36,12 +37,14 @@ CORE_CONFIG_YAML_FILE={{.CoreConfigYAMLFile}}
 	vars := struct {
 		LLMAPIKey          string
 		LLMBaseURL         string
+		LLMSimpleModel     string
 		LLMSDK             string
 		RootAPIBearerToken string
 		CoreConfigYAMLFile string
 	}{
 		LLMAPIKey:          config.LLMConfig.APIKey,
 		LLMBaseURL:         config.LLMConfig.BaseURL,
+		LLMSimpleModel:     config.LLMConfig.SimpleModel,
 		LLMSDK:             config.LLMConfig.SDK,
 		RootAPIBearerToken: config.RootAPIBearerToken,
 		CoreConfigYAMLFile: config.CoreConfigYAMLFile,
@@ -69,9 +72,10 @@ CORE_CONFIG_YAML_FILE={{.CoreConfigYAMLFile}}
 
 // LLMConfig contains LLM configuration
 type LLMConfig struct {
-	APIKey  string
-	BaseURL string
-	SDK     string
+	APIKey      string
+	BaseURL     string
+	SimpleModel string
+	SDK         string
 }
 
 // EnvConfig contains all environment configuration

@@ -38,9 +38,10 @@ func TestGenerateEnvFile(t *testing.T) {
 			// Create mock env config for testing
 			envConfig := &EnvConfig{
 				LLMConfig: &LLMConfig{
-					APIKey:  "test-api-key",
-					BaseURL: "https://api.example.com",
-					SDK:     "openai",
+					APIKey:      "test-api-key",
+					BaseURL:     "https://api.example.com",
+					SimpleModel: "test-model",
+					SDK:         "openai",
 				},
 				RootAPIBearerToken: "test-root-token",
 			}
@@ -63,19 +64,23 @@ func TestGenerateEnvFile(t *testing.T) {
 				// Check for required environment variables
 				assert.Contains(t, contentStr, "LLM_API_KEY=")
 				assert.Contains(t, contentStr, "LLM_BASE_URL=")
+				assert.Contains(t, contentStr, "LLM_SIMPLE_MODEL=")
 				assert.Contains(t, contentStr, "LLM_SDK=")
 				assert.Contains(t, contentStr, "ROOT_API_BEARER_TOKEN=")
 
 				// Verify config values are not empty
 				lines := strings.Split(contentStr, "\n")
-				var llmAPIKey, llmBaseURL, llmSDK, rootToken string
+				var llmAPIKey, llmBaseURL, llmSimpleModel, llmSDK, rootToken string
 				for _, line := range lines {
 					if strings.HasPrefix(line, "LLM_API_KEY=") {
 						llmAPIKey = strings.TrimPrefix(line, "LLM_API_KEY=")
 					}
 					if strings.HasPrefix(line, "LLM_BASE_URL=") {
 						llmBaseURL = strings.TrimPrefix(line, "LLM_BASE_URL=")
 					}
+					if strings.HasPrefix(line, "LLM_SIMPLE_MODEL=") {
+						llmSimpleModel = strings.TrimPrefix(line, "LLM_SIMPLE_MODEL=")
+					}
 					if strings.HasPrefix(line, "LLM_SDK=") {
 						llmSDK = strings.TrimPrefix(line, "LLM_SDK=")
 					}
@@ -85,10 +90,12 @@ func TestGenerateEnvFile(t *testing.T) {
 				}
 				assert.NotEmpty(t, llmAPIKey)
 				assert.NotEmpty(t, llmBaseURL)
+				assert.NotEmpty(t, llmSimpleModel)
 				assert.NotEmpty(t, llmSDK)
 				assert.NotEmpty(t, rootToken)
 				assert.Equal(t, "test-api-key", llmAPIKey)
 				assert.Equal(t, "https://api.example.com", llmBaseURL)
+				assert.Equal(t, "test-model", llmSimpleModel)
 				assert.Equal(t, "openai", llmSDK)
 				assert.Equal(t, "test-root-token", rootToken)
 			}

@@ -13,6 +13,7 @@
 FACTORIES: Mapping[str, COMPLETE_FUNC] = {
     "openai": openai_complete,
     "anthropic": anthropic_complete,
+    "minimax": openai_complete,
     "mock": mock_complete,
 }
 
@@ -61,7 +62,7 @@ async def llm_sanity_check():
 
 
 def response_to_sendable_message(message: LLMResponse) -> dict:
-    if DEFAULT_CORE_CONFIG.llm_sdk == "openai":
+    if DEFAULT_CORE_CONFIG.llm_sdk in ("openai", "minimax"):
         return message.raw_response.choices[0].message.model_dump()
     elif DEFAULT_CORE_CONFIG.llm_sdk == "anthropic":
         dp = {"role": message.role, "content": []}

@@ -21,7 +21,7 @@ class CoreConfig(BaseModel):
     llm_openai_default_header: Optional[Mapping[str, Any]] = None
     llm_openai_completion_kwargs: Mapping[str, Any] = {}
     llm_response_timeout: float = 60
-    llm_sdk: Literal["openai", "anthropic", "mock"] = "openai"
+    llm_sdk: Literal["openai", "anthropic", "minimax", "mock"] = "openai"
 
     llm_simple_model: str = "gpt-4.1"
 

@@ -0,0 +1,15 @@
+from acontext_core.llm.complete import FACTORIES
+from acontext_core.llm.complete.openai_sdk import openai_complete
+from acontext_core.schema.config import CoreConfig
+
+
+def test_minimax_uses_openai_compatible_adapter():
+    config = CoreConfig(
+        llm_api_key="test-key",
+        llm_base_url="https://api.minimax.io/v1",
+        llm_sdk="minimax",
+        llm_simple_model="MiniMax-M3",
+    )
+
+    assert config.llm_sdk == "minimax"
+    assert FACTORIES[config.llm_sdk] is openai_complete