8 AI tools researchers can actually trust
Tools that respect citations, support real workflows, and don't pretend to be smarter than they are. Graduate-student grade list.
Most "AI for research" lists confuse "tools that look impressive in a demo" with "tools that survive contact with a real literature review." Here are eight that actually earn their place.
1. SeekFiles AI — your personal literature
Upload your downloaded PDFs (papers, books, theses). Ask questions, get citation-grounded answers. Build separate Assistants per project so your Phase 1 lit-review doesn't leak into Phase 3.
2. Elicit — for finding new papers
Question-driven literature search across 200M+ papers. "What does the evidence say about X?" returns a structured table of papers with their key claims.
3. Connected Papers / Inciteful — for citation networks
Visualise how a seminal paper connects to predecessors and descendants. Great for understanding a field's structure.
4. Semantic Scholar — for actually finding papers
Free, ML-augmented search across academic literature. The Highlights feature surfaces relevant TLDRs without you reading every abstract.
5. Zotero (+ Better BibTeX) — for actual citation management
Not AI but the entire scholarly workflow runs on Zotero. AI tools above feed into Zotero. Don't skip this.
6. NotebookLM — for digesting one big source
When you've found the book that matters to your project, NotebookLM turns it into an audio overview + Q&A interface. Great for synthesising on a walk.
7. Otter.ai (or Whisper) — for interview transcription
Qualitative research lives or dies on transcript quality. Both are now good enough for first-pass; you still polish before coding.
8. Claude / ChatGPT — for writing-stuck moments
For drafting prose, getting unstuck on argumentation, or restructuring a section. Never for citations.
What we'd never recommend
- AI tools that "find citations for you" without showing the source. If you can't see the cited paper, the citation might not exist.
- "AI literature reviews" that promise to write the review for you. They produce confident-looking output that any reviewer will see through.
- Paid tools that lock you into their PDF library. Your research outlives any one tool's subscription — keep ownership of your PDFs.
A workflow we've seen work
- Discovery: Semantic Scholar / Elicit for finding papers.
- Triangulation: Connected Papers for citation structure.
- Capture: Zotero with the browser extension.
- Deep reading: NotebookLM for big books, SeekFiles for your downloaded library.
- Synthesis: Your own writing, with Claude / ChatGPT for friction-reduction.
- Citation: Zotero → Word / LaTeX / Google Docs.
The whole stack can run for free or close to it. The most expensive thing is a research-grade subscription to journals (if your library doesn't already cover it).
A note on integrity
The current academic norm on AI use is: disclose it, supervise it, never let it produce uncited claims. Your reviewers will be more sceptical of AI-assisted work in 2026 than they were in 2024. Lean into the supervision — the tools above support you, they don't replace you.
Like this? Get the next one in your inbox.
Weekly tips on getting more out of your file library — RAG, retrieval tricks, and product updates. No spam.
Try it free
Ask your files anything. Get answers with citations.
50 welcome credits. 3 assistants. No credit card. Upload your first file in under two minutes.