--- title: "deep-research vs storm" type: "comparison" canonical_url: "https://www.graphcanon.com/compare/dzhng-deep-research-vs-stanford-oval-storm" tools: ["dzhng-deep-research", "stanford-oval-storm"] --- # deep-research vs storm Neutral, constraint-first comparison with live GitHub stats. | | [deep-research](/tools/dzhng-deep-research.md) | [storm](/tools/stanford-oval-storm.md) | | --- | --- | --- | | Tagline | An AI-powered research assistant that performs iterative, deep research on any topic | An LLM-powered knowledge curation system that researches a topic and generates full-length reports with citations. | | Stars | 19,312 | 29,951 | | Forks | 1,973 | 2,802 | | Open issues | 90 | 123 | | Language | TypeScript | Python | | Adopt for | Deep Research is an AI-powered research assistant that efficiently dives into any topic by leveraging search engines, web scraping, and large language models. It iteratively refines its research direction over time to go | STORM is an LLM-powered knowledge curation system that uses agentic-RAG for deep research to generate full-length reports with citations. | | Persona | - | - | | Runtime | - | - | | License | MIT | MIT | | Categories | Data & Retrieval, AI Agents | Data & Retrieval, LLM Frameworks | ## Trust and health _Sourced signals - not a safety guarantee. No winner column._ | | [deep-research](/tools/dzhng-deep-research.md) | [storm](/tools/stanford-oval-storm.md) | | --- | --- | --- | | Maintenance | Steady (60%) | Slowing (36%) | | Days since push | 87d | 281d | | Open issues (now) | 90 | 123 | | Owner type | User | Organization | | Security scan | 39 low (39 low) | No criticals | | Full report | [trust report](/tools/dzhng-deep-research/trust.md) | [trust report](/tools/stanford-oval-storm/trust.md) | **Typed relationship:** deep-research _(alternative)_ storm Both aim at performing iterative and deep research using AI-powered methods, indicating a competitive relationship in the domain of AI-assisted research. ## Decision facts: deep-research - **Requirements:** Min 2 GB RAM; Requires Docker; Requires Node.js environment setup.; Needs specific API keys for third-party web search (Firecrawl) and language model services (OpenAI). - **Adopt for:** Deep Research is an AI-powered research assistant that efficiently dives into any topic by leveraging search engines, web scraping, and large language models. It iteratively refines its research direction over time to go ## Decision facts: storm - **Adopt for:** STORM is an LLM-powered knowledge curation system that uses agentic-RAG for deep research to generate full-length reports with citations. ## Choose when ### Choose deep-research if… - deep-research is primarily TypeScript; storm is Python. - Requirements: Min 2 GB RAM; Requires Docker; Requires Node.js environment setup.; Needs specific API keys for third-party web search (Firecrawl) and language model services (OpenAI).. - Both aim at performing iterative and deep research using AI-powered methods, indicating a competitive relationship in the domain of AI-assisted research. - Tags unique to deep-research: research, ai, o3-mini, gpt. - Also covers AI Agents. - deep-research ships Docker support for self-hosted deployment. - You require a detailed and comprehensive report on a specific topic where traditional manual or less refined automated tools are too basic. ### Choose storm if… - storm is primarily Python; deep-research is TypeScript. - Both aim at performing iterative and deep research using AI-powered methods, indicating a competitive relationship in the domain of AI-assisted research. - Tags unique to storm: large-language-models, report-generation, retrieval-augmented-generation, knowledge-curation. - Also covers LLM Frameworks. - When you need a tool capable of generating comprehensive and cited reports based on deep research. ## When NOT to use deep-research - If your project requires proprietary or classified data analysis because Deep Research relies on public web scraping and search engines, which limits access to non-public content. - You are looking for a tool that operates solely offline; since the tool needs internet access to perform its tasks through API calls to services like Firecrawl and OpenAI. ## When NOT to use storm - Avoid STORM if cost optimization is critical as it may involve using multiple different models to balance between quality and expense. - Do not choose STORM if you require a tool that does not modify its behavior through agentic-RAG processes, which are central to this system’s operation. ## Common questions ### What is the difference between deep-research and storm? deep-research: An AI-powered research assistant that performs iterative, deep research on any topic. storm: An LLM-powered knowledge curation system that researches a topic and generates full-length reports with citations.. See the comparison table for live GitHub stats and shared categories. ### When should I choose deep-research over storm? Choose deep-research over storm when deep-research is primarily TypeScript; storm is Python; Requirements: Min 2 GB RAM; Requires Docker; Requires Node.js environment setup.; Needs specific API keys for third-party web search (Firecrawl) and language model services (OpenAI).; Both aim at performing iterative and deep research using AI-powered methods, indicating a competitive relationship in the domain of AI-assisted research; Tags unique to deep-research: research, ai, o3-mini, gpt; Also covers AI Agents; deep-research ships Docker support for self-hosted deployment; You require a detailed and comprehensive report on a specific topic where traditional manual or less refined automated tools are too basic. ### When should I choose storm over deep-research? Choose storm over deep-research when storm is primarily Python; deep-research is TypeScript; Both aim at performing iterative and deep research using AI-powered methods, indicating a competitive relationship in the domain of AI-assisted research; Tags unique to storm: large-language-models, report-generation, retrieval-augmented-generation, knowledge-curation; Also covers LLM Frameworks; When you need a tool capable of generating comprehensive and cited reports based on deep research. ### When should I avoid deep-research? If your project requires proprietary or classified data analysis because Deep Research relies on public web scraping and search engines, which limits access to non-public content. You are looking for a tool that operates solely offline; since the tool needs internet access to perform its tasks through API calls to services like Firecrawl and OpenAI. ### When should I avoid storm? Avoid STORM if cost optimization is critical as it may involve using multiple different models to balance between quality and expense. Do not choose STORM if you require a tool that does not modify its behavior through agentic-RAG processes, which are central to this system’s operation. ### Is deep-research or storm more popular on GitHub? storm has more GitHub stars (29,951 vs 19,312). Stars measure visibility, not whether either tool fits your constraints. ### Are deep-research and storm open source? Yes - both are open-source projects on GitHub (deep-research: MIT, storm: MIT). ### Where can I find alternatives to deep-research or storm? GraphCanon lists graph-backed alternatives at /tools/dzhng-deep-research/alternatives and /tools/stanford-oval-storm/alternatives (/tools/dzhng-deep-research/alternatives.md, /tools/stanford-oval-storm/alternatives.md), ranked by typed relationship edges rather than popularity votes. ### Is there a machine-readable version of this comparison? Yes. The markdown twin at /compare/dzhng-deep-research-vs-stanford-oval-storm.md mirrors this page for agents and LLM crawlers, with the same stats table and FAQ answers. ### Which is better maintained, deep-research or storm? deep-research: Steady. storm: Slowing. Compare maintenance labels, days since push, and release cadence in the trust section below - stars alone do not measure maintenance. ### Where are the full trust reports for deep-research and storm? GraphCanon publishes per-repo trust reports with dated maintenance, provenance, and scan summaries: deep-research: /tools/dzhng-deep-research/trust; storm: /tools/stanford-oval-storm/trust. --- **Machine-readable endpoints** - JSON: [`/api/graphcanon/graph?tool=dzhng-deep-research`](/api/graphcanon/graph?tool=dzhng-deep-research) - LLM index: [/llms.txt](/llms.txt) - Full corpus: [/llms-full.txt](/llms-full.txt) _GraphCanon - The knowledge graph for AI development. https://www.graphcanon.com/_