Skip to main content
LocalREPL is the default execution environment for RLM. It runs Python code in the same process as the RLM host application, using Python’s built-in exec() function with a sandboxed namespace. The REPL shares the same virtual environment and memory space as the host process, but restricts access to dangerous builtins like eval, exec, and compile. This provides fast execution with minimal overhead, making it ideal for development and trusted code execution, but offers no process-level isolation from the host system. 
rlm = RLM(
    backend="openai",
    backend_kwargs={"model_name": "gpt-5-mini"},
    environment="local",
    environment_kwargs={
        "setup_code": "import json",
    },
)

​
Arguments

ArgumentTypeDefaultDescription
setup_codestrNoneCode to run at initialization
context_payloadstr | dict | listAutoInitial context (set by RLM)
lm_handler_addresstupleAutoSocket address (set by RLM)

​
How It Works

  1. Creates sandboxed globals with restricted __builtins__
  2. Injects context, llm_query(), llm_query_batched(), FINAL_VAR()
  3. Executes each code block via exec()
  4. llm_query() sends TCP requests to LM Handler
  5. Variables persist across code blocks in locals

​
Sandboxed Builtins

Allowed: print, len, range, str, int, float, list, dict, set, tuple, open, min, max, sum, sorted, enumerate, zip, map, filter, standard exceptions Blocked: eval, exec, compile, input, globals, locals

​
Limitations

  • Shares process memory with host
  • No network isolation
  • Dependencies must be installed in host virtualenv