Zero-Hassle AI Enabled Python Profiler for Time, Memory, return object size, cpu memory

AIPyMemTimeProfiler is a zero-configuration, AI-assisted Python profiler that automatically captures function-level memory and time usage — without any code changes.

Powered by LLMs, it not only measures performance but also analyzes your functions for inefficiencies and memory leaks.

Key Features

Zero code modifications — No decorators or wrappers needed

AI-powered analysis of selected functions

Function-level metrics for time, memory, arguments, return object size

Smart project detection — Profiles only your code, not external libraries

Structured JSON output for automation or visualization

Optional profiling of third-party libraries

Supports Flask apps, CLI scripts, nested folder hierarchies

Visualize Performance with Interactive Graphs

AIPyMemTimeProfiler includes a built-in Streamlit dashboard to visualize profiling data in a flamegraph-style format.

What You Get

• Multi-metric overview (time, memory, cpu time, return object size)

• Horizontal flame-style charts

• Interactive tooltips for every function

• No extra config — just run and view

Installation

pip install aipymemtimeprofiler

Quick Start

1. Set Environment Variables or pass arguments

export PROFILER_FILE_PATH="/absolute/path/to/your_script.py"
export PROFILER_DIR_PATH="/absolute/path/to/your/project/root"

If you just want to try, you can test it with sample_project.

export PROFILER_FILE_PATH="$(pwd)/sample_project/inside/app.py": The Python file to be profiled
export PROFILER_DIR_PATH="$(pwd)/sample_project/inside": Root of your project for accurate filtering

CLI entry point

profile_code

OR pass arguments

profile_code --file-path your_script.py --dir-path /your/output/dir

Arguments

| Argument | Env Variable | Description | Default | |----------------------------|----------------------------|-------------------------------------------------------|----------------------------------------------| | -h, --help | – | Show this help message and exit | – | | --file-path FILE_PATH | PROFILER_FILE_PATH | Path to the Python script to profile | None | | --dir-path DIR_PATH | PROFILER_DIR_PATH | Working directory and output path | ./sample_project/inside | | --console-display | CONSOLE_DISPLAY=True | Print profiling output to the console | False | | --agentic-profiler | AGENTIC_PROFILER=True | Enable AI-based interactive profiling assistant | False | | --ollama-model OLLAMA_MODEL | AGENT_NAME | LLM model name to use with the agentic profiler | deepseek-r1:1.5b | | --include-libraries | INCLUDE_LIBRARIES=True | Include external libraries during profiling | False | | --leak-threshold-kb LEAK_THRESHOLD_KB | – | Memory leak threshold in KB | 10 |

Launch the Graph UI

profile_graph

2. If you want to verify 3rd party libraries

export INCLUDE_LIBRARIES=True

LLM Environment Setup

Download Ollama from Ollama

ollama run <yout_model>

If you don't know which model to use.

ollama run deepseek-r1:1.5b: It is preferable as it is light weight.

Set your model env variable.

export AGENT_NAME="<your_model>"
export AGENTIC_PROFILER=True

TL;DR

pip install aipymemtimeprofiler export PROFILER_FILE_PATH=./app.py export PROFILER_DIR_PATH=./ profile_code

You’ll be prompted to select a function to analyze. Example:

| Index | Function | File Path | | ----- | ---------------- | ------------------------- | | 0 | calculate_data | /project/app/compute.py | | 1 | main | /project/app/main.py | | N | Skip Analysis | - |

Time & Memory Metrics

Automatically captures:

• Max execution time (CPU)
• Peak memory usage
• RSS memory growth
• Return object size
• Arguments passed

Project-Aware Analysis

Only your code is profiled.
System libraries and external modules are ignored using intelligent project-path detection.

Profiler Metrics

The following table describes the metrics collected by the profiler:

| Metric | Description | Key | |--------------------------|---------------------------------------------------------------------------------------------------|-----------------------------------| | Function Name | The name of the function being profiled. | function | | File Path | The absolute path to the file where the function is defined. | file | | Line Number | The line number where the function starts in the file. | line | | Execution Time | Maximum time taken by each function in milliseconds (ms). | max_time_ms | | CPU Time | Time spent by the CPU on this function (ms). | cpu_time_ms | | Peak Memory Usage | Maximum memory usage during function execution in kilobytes (KB). | max_mem | | RSS Memory Growth | Growth in Resident Set Size (RSS) memory in kilobytes (KB), helps spot memory leaks. | mem_growth_rss_kb | | Arguments | The arguments passed to the function being profiled. | args | | Possible Memory Leak | Indicates if a potential memory leak is detected (if any). | possible_memory_leak | | Notes | Any additional notes related to the profiling data. | note | | Returned Object Size | The size of the returned object in bytes. | return_obj |

Option to select a function for analysis, which is analysed by the Ollama model installed and configured. This is the table providing the options for analysis.

Available Functions for Analysis

| Index | Function Name | File Path | |-----------|------------------------|----------------------------------| | 0 | function_one | /path/to/file_one.py | | 1 | function_two | /path/to/file_two.py | | 2 | function_three | /path/to/file_three.py | | 3 | function_four | /path/to/file_four.py | | 4 | function_five | /path/to/file_five.py | | ... | ... | ... | | N | function_n | /path/to/file_n.py | | N+1 | Skip Analysis | - |

Structured JSON Reports

{ "function": "calculate_total", "file": "/app/logic.py", "line": 23, "cpu_time_ms": 15.3, "max_mem": 1024, "mem_growth_rss_kb": 300, "args": ["x", "y"], "return_obj": 124, "possible_memory_leak": false, "note": "" }

Works with Any Project Structure

Handles nested folder hierarchies easily — just point to your project root and go.

---

Supported Use Cases

• Pure Python projects
• Flask APIs and apps
• Any directory layout

---

Want More?

• [ ] Console table toggle
• [ ] HTML report output

Pull requests are welcome!