aim
A super-easy way to record, search and compare AI experiments.
Decision gist · record as of 2026-08-14
Yes, with conditions. Aim is suitable if you need a self-hosted, open-source experiment tracker with a polished UI and Python query API for ML teams. Install it if you can tolerate the medium dependency footprint and are willing to verify the four known security vulnerabilities do not affect your threat model. Avoid if you require active maintenance and rapid security patches—the 463-day gap since the last release signals aging support.AI-flagged interpretation of the facts on this page — verify before relying
Before you install
- Medium install friction due to 25 runtime dependencies including database (SQLAlchemy, alembic), web server (FastAPI, uvicorn), and specialized libraries (aimrocks, aimrecords).
- Maintenance status is aging—last release was 463 days ago—so expect slower updates and potential compatibility drift.
License · maintenance · safety
permissive license (permissive) — Apache 2.0 permissive license allows commercial and private use without restriction, though you must include a copy of the license and state any modifications.
last release 2025-05-08 (463 days)
4 known vulnerabilities (OSV.dev, 2026-08-14) · 192,207 downloads/mo, #9,866 on PyPI
Alternatives
Verify before relying
pip install aim
from aim import Run
run = Run()
run["loss"] = 0.5- Whether the web UI requires a separate server process or runs alongside the SDK
- Performance characteristics when handling thousands of concurrent runs as claimed
- Current status of the four known security vulnerabilities (GHSA-gmvv-rj92-9w35, GHSA-gp5h-f9c5-8355, PYSEC-2026-1087, PYSEC-2026-1088) and whether patches are available
What it is and what it does
Aim is an open-source ML experiment tracker designed to log training runs and arbitrary AI metadata, then make that data queryable and explorable. It provides both a web-based UI for visual comparison and filtering of runs, and a Python SDK for programmatic access to tracked metadata—useful for automation, analysis in Jupyter notebooks, or building custom dashboards. The package bundles a full web application stack (FastAPI, SQLAlchemy, Alembic) and specialized storage layers (aimrocks, aimrecords) to handle tracking of experiments.
The core workflow is straightforward: create a Run object, log metrics and parameters as you train, then browse or query results through the UI or API. Aim handles system monitoring, real-time alerting, and run organization via tags and experiments. It integrates with popular ML frameworks and supports migration from other trackers. However, the large dependency footprint and aging maintenance status mean you should verify compatibility with your Python version and check whether the known security vulnerabilities affect your use case.
Use it for
- Log and compare hyperparameter sweeps across training runs to identify the best configuration
- Track metrics, images, and system resource usage during model training to spot training issues in real time
- Query experiment metadata programmatically in a Jupyter notebook to analyze trends across runs
- Centralize experiment history for a team so all members can browse and filter runs by tags and parameters
- Migrate existing experiment logs from another tracker into Aim's queryable format for unified analysis
Worth the install?
AI-flagged interpretation of the facts on this page. Verify before relying on it.
Yes, with conditions.
Aim is suitable if you need a self-hosted, open-source experiment tracker with a polished UI and Python query API for ML teams. Install it if you can tolerate the medium dependency footprint and are willing to verify the four known security vulnerabilities do not affect your threat model. Avoid if you require active maintenance and rapid security patches—the 463-day gap since the last release signals aging support.
Install
aim on PyPI
Before you install
Medium install friction due to 25 runtime dependencies including database (SQLAlchemy, alembic), web server (FastAPI, uvicorn), and specialized libraries (aimrocks, aimrecords). Maintenance status is aging—last release was 463 days ago—so expect slower updates and potential compatibility drift.
License in practice
Apache 2.0 permissive license allows commercial and private use without restriction, though you must include a copy of the license and state any modifications.
Quickstart
pip install aim
from aim import Run
run = Run()
run["loss"] = 0.5Verify before relying
- Whether the web UI requires a separate server process or runs alongside the SDK
- Performance characteristics when handling thousands of concurrent runs as claimed
- Current status of the four known security vulnerabilities (GHSA-gmvv-rj92-9w35, GHSA-gp5h-f9c5-8355, PYSEC-2026-1087, PYSEC-2026-1088) and whether patches are available
Package facts
| License | permissive license permissive |
| Python support | Supports the current Python release >=3.7.0 |
| Install friction | Medium. Platform-specific wheel |
| Runtime dependencies | 25 packagesaim-uiaimrecordsaimrockscachetoolsclickcryptographyfilelocknumpypsutilRestrictedPythontqdmaiofilesalembicfastapijinja2pytzSQLAlchemyuvicornPillowpackagingpython-dateutilrequestswatchdogwebsocketsboto3 |
| Maintenance | Aging 463 days since the last release |
| First released | |
| Downloads | 192,207 / month, #9,866 on PyPI 30-day window, as of 2026-08-14 |
| Known vulnerabilities | 4 GHSA-gmvv-rj92-9w35, GHSA-gp5h-f9c5-8355, PYSEC-2026-1087, PYSEC-2026-1088 |
| Classifiers | License :: OSI Approved :: Apache Software LicenseProgramming Language :: PythonProgramming Language :: Python :: 3Programming Language :: Python :: 3.10Programming Language :: Python :: 3.11Programming Language :: Python :: 3.12Programming Language :: Python :: 3.7Programming Language :: Python :: 3.8Programming Language :: Python :: 3.9Programming Language :: Python :: Implementation :: PyPy |
Evidence: aim-3.29.1-cp310-cp310-macosx_10_14_x86_64.whl; aim-3.29.1-cp310-cp310-macosx_11_0_arm64.whl; aim-3.29.1-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl; aim-3.29.1-cp310-cp310-manylinux_2_24_x86_64.whl; aim-3.29.1-cp310-cp310-manylinux_2_28_x86_64.whl; aim-3.29.1-cp311-cp311-macosx_10_14_x86_64.whl; aim-3.29.1-cp311-cp311-macosx_11_0_arm64.whl; aim-3.29.1-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl; aim-3.29.1-cp311-cp311-manylinux_2_24_x86_64.whl; aim-3.29.1-cp311-cp311-manylinux_2_28_x86_64.whl; aim-3.29.1-cp312-cp312-macosx_10_14_x86_64.whl; aim-3.29.1-cp312-cp312-macosx_11_0_arm64.whl; aim-3.29.1-cp312-cp312-manylinux_2_28_x86_64.whl; aim-3.29.1-cp37-cp37m-macosx_10_14_x86_64.whl; aim-3.29.1-cp37-cp37m-manylinux_2_24_x86_64.whl; aim-3.29.1-cp38-cp38-macosx_10_14_x86_64.whl; aim-3.29.1-cp38-cp38-macosx_11_0_arm64.whl; aim-3.29.1-cp38-cp38-manylinux_2_17_x86_64.manylinux2014_x86_64.whl; aim-3.29.1-cp38-cp38-manylinux_2_24_x86_64.whl; aim-3.29.1-cp38-cp38-manylinux_2_28_x86_64.whl
Tags
Let your AI agent find packages like this
Example. Real query, live index.
You found this page by searching. An agent finds it by wishing: SkillFed indexes 14,416 PyPI packages by what they can do, searchable in plain language.
wish › “ml metadata visualization”
- aimAim logs training runs and AI metadata, then provides a web UI and…
- tracemlTraceML tracks metrics, parameters, artifacts, and data references…
- ai-edge-model-explorerVisualizes and debugs machine learning model graphs with hierarchical…
Give your agent the search over MCP, or paste the wish link into any chat.
More Artificial Intelligence packages
LiteLLM provides a unified Python interface to call 100+ LLM providers (OpenAI, Anthropic, Gemini, Bedrock, Azure, and others) using OpenAI-compatible API format, available as both a Python SDK and a self-hosted AI Gateway proxy server.
Install it if you need to work with multiple LLM providers or want to centralize LLM routing in your organization.
Client library and CLI tool for downloading, uploading, and managing models, datasets, and repositories on the Hugging Face Hub platform.
Install it if you work with Hugging Face Hub models or datasets.
LangChain provides a framework for building agents and LLM-powered applications by composing language models, tools, and memory through a unified API that abstracts over multiple model providers.
hf-xet provides chunk-based deduplication and efficient file transfer for the Hugging Face Hub, enabling faster uploads and downloads of large files with local disk caching.
Tokenizers converts raw text into token sequences for NLP models, with support for training custom vocabularies and using pre-built tokenizers (BPE, WordPiece) optimized for speed via Rust.
Transformers provides a unified framework for loading, fine-tuning, and running state-of-the-art pretrained models across text, vision, audio, video, and multimodal tasks using PyTorch, JAX, or TensorFlow.
Install it if you need to run or train any transformer-based model for NLP, vision, audio, or multimodal tasks.
See also comet-ml · isa-rwval · sagemaker-experiments · azureml-mlflow · trackio · whylogs · wandb · azureml-telemetry · traceml · mlflow