Alibi
A Python library for explaining machine learning models.
- Status
- Not yet determined
- Last checked
- Checked · Oct 5, 2026
- License
- Open source
- Pricing
- Free tier
What it is
Alibi is an open-source Python library designed to provide algorithms for explaining machine learning models. It focuses on interpretability and explainability, offering tools for counterfactual explanations and other interpretability techniques. Alibi is particularly useful for developers and data scientists who need to understand and trust the decisions made by complex models. Despite its utility, the library has a relatively high number of open issues on GitHub, indicating potential areas for improvement or active development.
Focus on counterfactual explanations for model interpretability.
When it fits — and when it doesn’t
A good fit for
Developers needing model interpretability
Alibi provides tools for understanding complex model decisions.
Data scientists working with AI trust issues
The library's focus on explainability helps build trust in AI systems.
Look elsewhere if
Projects requiring frequent updates
The library has not seen frequent updates recently.
Users needing extensive documentation
Lack of a dedicated homepage or repository link might limit available resources.
Known weaknesses
Each one links to the public issue, changelog or docs page it came from, where we have one.
- 01
High number of open issues
159 open issues on GitHub
- 02
Limited recent updates
Last commit was in October 2025
- 03
No official homepage or repository link on npm
Missing homepage and repository fields in npm data
What it does well
High GitHub stars
2643 stars on GitHub
Wide range of explanation techniques
Supports counterfactual explanations and other interpretability methods
Active contributor base
25 contributors on GitHub
Pricing
- Free tier
- Yes
- Paid from
- —
- Model
- N/A
- Enterprise plan
- No
Pricing last confirmed Sep 27, 2026.
Performance benchmarks
How Fast Is It?
Set up Alibi
Setup steps, common gotchas and what to check after 30 days.