Top AI Repos — open-source AI, indexed and scored
Top AI Repos tracks AI repositories on GitHub and answers two different questions about each one: is it moving right now, and would you bet a product on it.
Top AI Repos tracks AI repositories on GitHub and answers two different questions about each one: is it moving right now, and would you bet a product on it.
Kura is a simple reproduction of the CLIO paper which uses language models to label user behaviour before clustering them based on embeddings recursively. This helps us understand user behaviour on a higher level without sacrificing PII.
| Date | Stars |
|---|---|
| 2026-07-31 | 510 |
| 2026-08-05 | 511 |
| 2026-08-06 | 511 |
Today
— stars today
This week
— stars this week
This month
— stars this month
Momentum
0.0
growth rate 0.00%/day
# Kura: Procedural API for Chat Data Analysis  [](https://pypi.org/project/kura/) [](https://github.com/567-labs/kura/stargazers) [](https://567-labs.github.io/kura/) [](https://opensource.org/licenses/MIT) [](https://pypi.org/project/kura/) [](https://pypi.org/project/kura/) **Your AI assistant handles thousands of conversations daily. But do you know what users actually need?** Kura is an open-source library for understanding chat data through machine learning, inspired by [Anthropic's CLIO](https://www.anthropic.com/research/clio). It automatically clusters conversations to reveal patterns, pain points, and opportunities hidden in your data. ## The Hidden Cost of Not Understanding Your Users Every day, your AI assistant or chatbot has thousands of conversations. Within this data lies critical intelligence: - **80% of support tickets** might stem from the same 5 unclear features - **Key feature requests** repeated by hundreds of users in different ways - **Revenue opportunities** from unmet needs you didn't know existed - **Critical failures** affecting user trust that go unreported Manually reviewing conversations doesn't scale. Traditional analytics miss semantic meaning. **Kura bridges this gap.** ## What Kura Does Kura transforms unstructured conversation data into structured insights: ``` 10,000 conversations → AI Analysis → 20 clear patterns ``` - **Automatic Intent Discovery**: Find what users actually want (not what they say) - **Failure Pattern Detection**: Identify where your AI falls short before users complain - **Feature Priority Insights**: See which missing features impact the most users - **Semantic Clustering**: Group by meaning, not keywords - **Privacy-First Design**: Analyze patterns without exposing individual conversations ## Real-World Impact ### E-commerce Support Bot **Challenge**: 50,000 weekly conversations, unknown pain points **Discovery**: 35% of conversations about shipping clustered into 3 issues **Result**: Fixed root causes, reduced support volume by 40% ### Developer Documentation Assistant **Challenge**: Users struggling but not reporting specific issues **Discovery**: 2,000+ conversations revealed 5 consistently confusing APIs **Result**: Targeted doc improvements, 60% reduction in those queries ### SaaS Onboarding Bot **Challenge**: 30% of trials not converting, unclear why **Discovery**: Clustering revealed 3 missing integration requests **Result**: Built integrations, trial conversion increased 18% ## Installation ```bash uv pip install kura ``` ## When to Use Kura **Kura is perfect when you have:** - 100+ conversations to analyze (scales to millions) - A need to understand user patterns, not individual conversations - Unstructured conversation data from chatbots, support systems, or AI assistants - Questions like "What are users struggling with?" or "What features are they requesting?" **Kura might not be the best fit if:** - You have fewer than 100 conversations (manual review might be faster) - You need real-time analysis (Kura is designed for batch processing) - You only need keyword-based search (use traditional search tools instead) - You require conversation-level sentiment analysis (Kura focuses on patterns) ## Common Use Cases ### Product Teams - **Feature Discovery**: Find the features users ask for in their own words - **Pa
Excerpt of 8,554 characters
Read on GitHub
Jason Liu · OpenAI · United States
140
109
6
5
1
Would you bet a product on this? Bounded 0–100 and slow moving.
matched fp:71b4ef0c5eb2d3ad, llm:Description: 'Kura is a simple reproduction of the CLIO paper which uses language models to label user behaviour before clustering them based on embeddings recursively. This helps us understand user behaviour on a higher level without sacrificing PII.' Language: Python.
matched fp:71b4ef0c5eb2d3ad, llm:Description: 'Kura is a simple reproduction of the CLIO paper which uses language models to label user behaviour before clustering them based on embeddings recursively. This helps us understand user behaviour on a higher level without sacrificing PII.' Language: Python.
matched fp:71b4ef0c5eb2d3ad, llm:Description: 'Kura is a simple reproduction of the CLIO paper which uses language models to label user behaviour before clustering them based on embeddings recursively. This helps us understand user behaviour on a higher level without sacrificing PII.' Language: Python.
matched fp:71b4ef0c5eb2d3ad, llm:Description: 'Kura is a simple reproduction of the CLIO paper which uses language models to label user behaviour before clustering them based on embeddings recursively. This helps us understand user behaviour on a higher level without sacrificing PII.' Language: Python.