STORM changelog: what's new each month
Every stable STORM release summarised by month: the highlights, new features, improvements, fixes and anything you need to act on. 3 months covered; the current month updates daily.
January 2025
1 release: 1.1.0
January brought expanded model support and new search capabilities to STORM. The release adds flexibility for users working with different language model providers.
Highlights
- STORM now works with any language model supported by LiteLLM, greatly expanding model options.
- Azure AI Search is now available as a retrieval option.
- Fixed a bug in the Serper retriever.
New
- Support for LiteLLM-compatible language models in STORM and Co-STORM.
- Azure AI Search retriever integration.
Fixed
- Serper retriever bug fix.
September 2024
1 release: 1.0.0
September brought the full launch of Co-STORM, a new collaborative variant of STORM that integrates into the knowledge-storm Python package. The approach was also accepted to EMNLP 2024.
Highlights
- Co-STORM is now fully integrated into the knowledge-storm package and ready to use.
- New Agent interface provides a unified framework for defining different LM agent policies.
- Users can now inject their own utterances to steer conversations interactively.
- Moderator Agent guides discussions by generating insightful questions and highlighting underexplored areas.
- Co-STORM LLM Experts ground responses in external knowledge sources and ask follow-up questions.
New
- Agent interface for defining LM agent policies for information seeking and knowledge curation
- Co-STORM LLM Experts agents that ground responses in external knowledge and generate follow-up questions
- Moderator Agent that guides conversation with insightful questions
- Human Engagement feature allowing users to inject utterances and steer conversations
- Co-STORM collaborative knowledge curation system
August 2024
1 release: 0.2.4
August brought expanded model support and improved retrieval capabilities to STORM. The release added new AI model integrations and enhanced the vector retrieval system for better performance.
Highlights
- Added support for Google and DeepSeek models
- Introduced Serper and Brave as new retrieval modules
- Decoupled vector store construction from the retriever for more flexible configuration
New
- GoogleModel for LLM tasks
- DeepSeekModel for LLM tasks
- SerperRM retrieval module
- BraveRM retrieval module
Improved
- VectorRM now decouples vector store construction from the retriever class for better modularity
Summaries are written automatically from the official release notes (full changelog ↗); check the original notes before relying on a detail. STORM: pricing, features and alternatives · All changelogs