Troubleshooting
🚧 Cortex.cpp is currently under development. Our documentation outlines the intended behavior of Cortex, which may not yet be fully implemented in the codebase.
Model No Response​
In this issue, the model and the engine load successfully, but when interacting with the model, the model either times out or does not respond. This issue may occur when Cortex is running with low or insufficient VRAM hardware.
To resolve this issue:
- Windows
- Mac
cortex stop
to stop the Cortex process and model if you have started a model before.- To make sure the processes are terminated, open your Task Manager.
- Search for any Jan or Cortex processes → Terminate the tasks.
- Rerun
cortex
.
cortex stop
to stop the Cortex process and model if you have started a model before.- To make sure the processes are terminated, open your Activity Monitor.
- Search for any Jan or Cortex processes → Terminate the tasks.
- Rerun
cortex
Axios Error​
AxiosError: getaddrinfo ENOTFOUND api.github.com
This issue occurs for several reasons:
- Bad internet connection when Cortex trying to download the model or the engine data.
- If you attempt to run a model with a specific engine that has not been initialized yet after installing Cortex.
To resolve this issue:
- Ensure your internet connection is good, or restart your model.
- Follow the command below to restart the Cortex process:
# Stop the Cortex processcortex stop# Start the Cortex processcortex# Init the enginecortex engines <engine_name> init
Engine Installation Error​
Installing engine . . .Error:EPERM: operation not permitted
This issue occurs when an engine runs in the background and you attempt to run a different engine without stopping the previous one.
To resolve this issue, follow the steps below:
# Stop the Cortex processcortex stop# Start the Cortex processcortex# Init the enginecortex run <model_id>:<engine_name>
500 Status Code​
"500 status code ("no body") indicates an internal server error with no additional details provided. This issue may occur for several reasons:
- Cortex is running in low or insufficient VRAM hardware.
- The engine version is outdated, while the model requires the latest version of the engine.
To resolve this issue:
- Windows
- Mac
To solve the insufficient VRAM:
cortex stop
to stop the Cortex process and model if you have started a model before.- To make sure the processes are terminated, open your Task Manager.
- Search for any Jan or Cortex processes → Terminate the tasks.
- Rerun
cortex
.
To solve the error caused by the outdated engine version:
- Re-initialize the engine that you want to use:
# ONNX Enginecortex engines onnx init# TensorRT-LLM Enginecortex engines tensorrt-llm init
- Rerun
cortex
→ run the model you want to use.
To solve the insufficient VRAM:
cortex stop
to stop the Cortex process and model if you have started a model before.- To make sure the processes are terminated, open your Activity Monitor.
- Search for any Jan or Cortex processes → Terminate the tasks.
- Rerun
cortex
.
To solve the error caused by the outdated engine version:
- Re-initialize the engine that you want to use:
# ONNX Enginecortex engines onnx init# TensorRT-LLM Enginecortex engines tensorrt-llm init
- Rerun
cortex
→ run the model you want to use.
426 Error Code​
This issue occurs when a new Cortex process is started without fully stopping the previous Cortex processes. This causes a heavy load on port 39281
and requires a protocol upgrade.
To resolve this issue:
- Windows
- Mac
# Check for network connections using port 1337 to find the Cortex PID.netstat -an | findstr :1337# Terminate Cortex process by its Process ID (PID).taskkill /f /pid [pid]
# Check for network connections using port 1337 to find the Cortex PID.lsof -i :1337# Terminate Cortex process by its Process ID (PID).kill -9 [pid]