trackio
A lightweight, local-first, and free experiment tracking library built on top of Hugging Face Datasets and Spaces.
Decision gist · record as of 2026-08-14
Yes. Trackio is actively maintained, has no known vulnerabilities, installs with low friction, and offers a genuinely useful local-first alternative to cloud-dependent trackers. The wandb API compatibility makes adoption frictionless. MIT licensing imposes no restrictions. It is well-suited for researchers, agents, and teams who want experiment tracking without signup or infrastructure overhead.AI-flagged interpretation of the facts on this page — verify before relying
Before you install
- Requires Python 3.10 or higher.
- Low friction: pure Python wheel with no compiled dependencies.
- Active maintenance—released 1 day ago with 1642 GitHub stars.
License · maintenance · safety
permissive license (permissive) — MIT license (permissive): you can use, modify, and distribute Trackio freely in commercial and private projects without restriction.
last release 2026-08-13 (1 days) · last repo commit 2026-08-13 · 1,642 stars
0 known vulnerabilities (OSV.dev, 2026-08-14) · 606,512 downloads/mo, #5,790 on PyPI
Alternatives
Verify before relying
pip install trackio
import trackio
trackio.init(project="my-project", config={"lr": 0.001})
trackio.log({"loss": 0.5})
trackio.finish()
trackio.show()- Throughput limits for parallel experiments and SQLite write concurrency under real workloads.
- Performance characteristics when querying large experiment databases via CLI.
- Compatibility with custom frontends beyond the minimal starter template.
- Behavior and data consistency when syncing offline projects to Spaces with concurrent writes.
What it is and what it does
Trackio is a lightweight experiment tracking library designed for machine learning workflows and autonomous agents. It stores experiment metadata, metrics, and media in a local SQLite database, avoiding the need for account creation or cloud signup. The library provides a wandb-compatible API (init, log, finish), so existing logging code can often be used without modification by importing trackio as wandb. It includes a Gradio-based web dashboard for viewing results, a CLI for querying experiment data with SQL, and optional integration with Hugging Face Spaces for collaborative sharing—all free.
The package depends on a modest set of runtime libraries: starlette and uvicorn for the dashboard server, gradio-client and huggingface-hub for Space integration, numpy and pillow for media handling, orjson for serialization, brotli for compression, python-multipart for form parsing, and tomli for config. It is designed to handle high-throughput parallel logging and to be LLM-friendly, with programmatic access to run management and SQL-based data querying for autonomous analysis.
Use it for
- Log metrics from parallel training runs locally without cloud infrastructure or authentication.
- Replace wandb in existing scripts by importing trackio as wandb, keeping logging code unchanged.
- Query experiment data programmatically via CLI SQL interface for LLM-driven analysis and decision-making.
- Share experiment results by syncing local projects to a free Hugging Face Space for team collaboration.
- Embed live experiment dashboards on websites or blogs using Space URLs with query parameter filtering.
- Track autonomous agent experiments with high-throughput logging and direct database access for inspection.
Worth the install?
AI-flagged interpretation of the facts on this page. Verify before relying on it.
Yes.
Trackio is actively maintained, has no known vulnerabilities, installs with low friction, and offers a genuinely useful local-first alternative to cloud-dependent trackers. The wandb API compatibility makes adoption frictionless. MIT licensing imposes no restrictions. It is well-suited for researchers, agents, and teams who want experiment tracking without signup or infrastructure overhead.
Install
trackio on PyPI
Before you install
Low friction: pure Python wheel with no compiled dependencies. Active maintenance—released 1 day ago with 1642 GitHub stars. Requires Python 3.10 or higher.
Requires Python 3.10 or higher.
License in practice
MIT license (permissive): you can use, modify, and distribute Trackio freely in commercial and private projects without restriction.
Quickstart
pip install trackio
import trackio
trackio.init(project="my-project", config={"lr": 0.001})
trackio.log({"loss": 0.5})
trackio.finish()
trackio.show()Verify before relying
- Throughput limits for parallel experiments and SQLite write concurrency under real workloads.
- Performance characteristics when querying large experiment databases via CLI.
- Compatibility with custom frontends beyond the minimal starter template.
- Behavior and data consistency when syncing offline projects to Spaces with concurrent writes.
Package facts
| License | permissive license permissive |
| Python support | Supports the current Python release >=3.10 |
| Install friction | Low. Pure-Python wheel |
| Runtime dependencies | 10 packagesbrotligradio-clienthuggingface-hubnumpyorjsonpillowpython-multipartstarlettetomliuvicorn |
| Maintenance | Actively maintained 1 days since the last release |
| Last repo commit | |
| First released | |
| Downloads | 606,512 / month, #5,790 on PyPI 30-day window, as of 2026-08-14 |
| Known vulnerabilities | None known OSV.dev, checked 2026-08-14 |
| Classifiers | License :: OSI Approved :: MIT LicenseOperating System :: OS IndependentProgramming Language :: Python :: 3 |
Evidence: trackio-0.35.0-py3-none-any.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 › “experiment tracking library”
- trackioTrackio is a local-first experiment tracking library that logs…
- dvc-studio-clientClient library for posting experiment metrics, model registry data,…
- neptuneNeptune is a Python client for experiment tracking and ML metadata…
Give your agent the search over MCP, or paste the wish link into any chat.
More Monitoring packages
Wraps any iterable to display a real-time progress bar in the terminal or Jupyter notebook, showing iteration count, elapsed time, and estimated time remaining.
Provides generated Python code for OpenTelemetry semantic conventions, enabling standardized attribute naming and constant definitions for instrumentation and telemetry collection.
Install it if you are using OpenTelemetry and want to follow semantic conventions correctly.
Provides the reference implementation of the OpenTelemetry API for collecting and exporting traces, metrics, and logs from Python applications.
Provides the abstract API and interfaces for OpenTelemetry instrumentation in Python, defining how to emit traces, metrics, and logs without tying code to a specific SDK implementation.
Exports OpenTelemetry observability data to an OpenTelemetry Collector using Protobuf-encoded messages over HTTP.
Install it if you are using OpenTelemetry in Python and need to send data to a Collector over HTTP.
Provides automatic instrumentation commands and programmatic APIs to inject distributed tracing into Python applications without code changes, detecting and instrumenting packages used by your program.
Install it if you need distributed tracing without code changes and have compatible instrumented packages in your environment.
See also wandb · aim · comet-ml · traceml · dvc-studio-client · dvclive · azureml-mlflow · neptune-scale · mlflow-skinny · valohai-utils