Research & Search
SaaS, open source and skills for research & search work — with the limits of each written down.Before you pick
What decides it
Reach decides the answer, not writing quality: SEO-ranked search results, a closed scholarly index, or the paywalled material most of them cannot open.
What to avoid
A citation that resolves and looks on-topic often does not support the sentence it is attached to; support falls as a run gathers pages.
What is changing
Source authority and citation checking are moving out of the model into the plumbing: curated authority indexes before retrieval, automated link verification after drafting.
SaaS
Hosted products you sign up for.Leading
Established picks, at the top of today's aggregate.
- 1OpenAI Deep ResearchBest overall for breadth and report quality — runs a multi-step autonomous loop reading full pages and citing sources.
- 2Perplexity (Deep Research)Fast citation-backed answers with a generous free tier — the go-to for quick sourced web research.
- 3Claude ResearchBest for analytical depth and reasoning over sources when you want careful synthesis over sheer breadth.
- 4Gemini Deep ResearchStrong long-form research reports with a meaningful free tier and deep Google ecosystem integration.
- 5ElicitAcademic literature review across millions of papers — extraction and screening for systematic reviews.
Emerging
Newer challengers, surfaced by two or more sources.
Open source
Repositories you run yourself.Leading
- 1GPT ResearcherThe recognized open-source deep-research agent (28k★) — plans, searches, reads many sources, and writes a long, cited report; ships as a Python package and full web app.
- 2STORM (Stanford OVAL)Stanford's LLM knowledge-curation system (30.3k★) — researches a topic and generates a full-length, Wikipedia-style report with citations; Co-STORM adds human-in-the-loop.
Emerging
- 1nanoAgentTeam/research-clawSelf-hosted Python agent for academic work: manages LaTeX and Overleaf projects, searches arXiv and PubMed, tracks conference deadlines, reachable from Telegram or Feishu.
- 2dralkh/seeraiZotero 9 plugin that chats with your library, searches eleven scholarly databases at once, runs OCR, and drives PRISMA systematic reviews.
- 3zoharbabin/web-researcher-mcpMCP server giving Claude or Cursor 37 research tools, with domain allowlists and Crossref checks that catch dead links and retractions.
- 4blurryface13/asteria-agentSelf-hosted deep research app that splits a question into sub-queries, searches the web, and exports cited reports; FastAPI, LangGraph, Next.js, per-user Postgres.
- 5hxcm-cre/scholar-agentConversational literature assistant that searches arXiv and your Zotero library, reads full texts in a side panel, and numbers citations automatically.
Skills
Skills you drop into a coding agent.Leading
- 1last30daysThe most-starred research skill (53.4k★, MIT, reached #1 on GitHub Trending) — sweeps Reddit, X, YouTube, HN, and the web for a topic and synthesizes a grounded, cited summary.
- 2imbad0202/academic-research-skillsA full research -> write -> review -> revise pipeline skill (39.3k★) with a multi-agent research team, systematic-review modes, and citation verification.
- 3claude-deep-research-skillAn 8-phase deep-research skill with source-credibility scoring and automated citation/DOI verification for long reports. Smaller/newer, so ranked mid at medium confidence.
- 4lingzhi227/agent-research-skillsA systematic academic literature-review skill running a 6-phase survey from frontier scan to compiled report. Niche and low-star, so ranked last.
Emerging
- 1K-Dense-AI/scientific-agent-skillsAdds 158 science skills to a coding agent — bioinformatics, chemistry, drug discovery — with connectors to PubChem, ChEMBL and 100 other databases.
- 2Orchestra-Research/AI-Research-SKILLs98 skills that walk a coding agent through ML research: literature survey, experiment design, training runs, evaluation and paper drafting.
- 3anysearch-ai/anysearch-skillOne API skill giving an agent web and vertical search, parallel batch queries and full-page extraction; works anonymously or with a free key.
- 4Socialpranker/deepdiveClaude Code skill that runs research as an audited pipeline: plan gate, parallel sub-agent search, red-team pass, and link-and-quote citation checks.
- 5hint-shu/deep-researchSix Claude Code skills laddered by effort, from a one-minute fact check to an hour-long synthesis, over Firecrawl, Tavily and Exa.