Skip to main content
POST
/
configure
Configure MemOS
curl --request POST \
  --url https://api.example.com/configure \
  --header 'Content-Type: application/json' \
  --data '
{
  "model_schema": "NOT_SET",
  "user_id": "root",
  "session_id": "0ce84b9c-0615-4b9d-83dd-fba50537d5d3",
  "chat_model": {
    "backend": "<string>",
    "config": {},
    "model_schema": "NOT_SET"
  },
  "mem_reader": {
    "backend": "<string>",
    "config": {},
    "model_schema": "NOT_SET"
  },
  "mem_scheduler": {
    "backend": "<string>",
    "config": {},
    "model_schema": "NOT_SET"
  },
  "user_manager": {
    "backend": "sqlite",
    "config": {}
  },
  "max_turns_window": 15,
  "top_k": 5,
  "enable_textual_memory": true,
  "enable_activation_memory": false,
  "enable_parametric_memory": false,
  "enable_mem_scheduler": false,
  "PRO_MODE": false
}
'
{
  "message": "Operation successful",
  "code": 200,
  "data": "<unknown>"
}

Body

application/json
​
model_schema
string
default:NOT_SET

Schema for configuration. This value will be automatically set.

​
user_id
string
default:root

User ID for the MOS. This is used to distinguish between different users' memories.

​
session_id
string
default:0ce84b9c-0615-4b9d-83dd-fba50537d5d3

Session ID for the MOS. This is used to distinguish between different dialogue

​
chat_model
LLMConfigFactory · object

Factory class for creating LLM configurations.

​
mem_reader
MemReaderConfigFactory · object

Factory class for creating MemReader configurations.

​
mem_scheduler
SchedulerConfigFactory · object

Memory scheduler configuration for managing memory operations

​
user_manager
UserManagerConfigFactory · object

Factory for user manager configurations.

​
max_turns_window
integer
default:15

Maximum number of turns to keep in the conversation history

​
top_k
integer
default:5

Maximum number of memories to retrieve for each query

​
enable_textual_memory
boolean
default:true

Enable textual memory for the MemChat

​
enable_activation_memory
boolean
default:false

Enable activation memory for the MemChat

​
enable_parametric_memory
boolean
default:false

Enable parametric memory for the MemChat

​
enable_mem_scheduler
boolean
default:false

Enable memory scheduler for automated memory management

​
PRO_MODE
boolean
default:false

Enable PRO mode for complex query decomposition

Response

Successful Response

Response model for configuration endpoint.

​
message
string
required

Response message

Example:

"Operation successful"

​
code
integer
default:200

Response status code

Example:

200

​
data
null

Response data