aitrove. Submit a tool

Best AI Research Assistant Tools

7 AI tools that fit the work finding, digesting and citing sources actually do, grouped by task. Tasks are ordered by how central they are to the job; inside each task, tools are ordered by aitrove popularity. That's not a quality ranking: we compiled every review from public sources (pricing pages, docs, published reviews, user reports), not hands-on testing.

Research workspaces · Terminal and agent research · Specialized research

Quick answer

The most popular pick on aitrove for each part of the job (popularity, not a quality score):

At a glance

ToolTaskBest forBiggest weaknessFree tierStarting price
NotebookLM (Gemini Notebook)Research workspacesStudents and researchers digesting a pile of documents.Usage is compute-based, refreshing every 5 hours up to a weekly limitYesFree tier: yes. Higher limits via Google AI Pro ($19.99/mo). (checked Oct 10, 2026)
Masala (formerly Panel)Research workspacesResearchers wanting an agent workspace.Repo renamed from Panel to MasalaYesFree and open source (MIT) (checked Oct 14, 2026)
InnogathResearch workspacesResearchers wanting cited AI reports.Citations still need checkingNot statedPricing not verified (checked Oct 14, 2026)
Perplexity Spaces (now Projects)Research workspacesTeams running ongoing research topics in Perplexity.Spaces have been renamed Projects, so the listing name is staleNot statedIncluded with Perplexity. Collaboration on Pro ($20/mo) and up. (checked Oct 11, 2026)
rStationTerminal and agent researchResearchers and hobbyists who want to point their agents at open math and CS problems.Reviews are automated; human verification of 'solved' problems is recorded separatelyFree to browseFree to browse and join; you supply your own model and compute. (checked Oct 19, 2026)
MoleTerminal and agent researchDevelopers wanting controlled research agents.YoungYesFree and open source (Apache-2.0) (checked Oct 14, 2026)
QED ScienceSpecialized researchAcademic scientists who want a critical pre-submission review of a paper or grant.Its FAQ says users grant QED a license to optionally train and evaluate its own models on uploads, and learned parameters remain after you delete a projectYesFree for academic researchers; Premium/Validity Machine pricing not published (early access). (checked Oct 13, 2026)

"Free tier" comes from each tool's pricing line; "Not stated" means our source didn't say. Prices include the date we checked them.

Research workspaces

1. NotebookLM (Gemini Notebook)

Best for: Students and researchers digesting a pile of documents.

Pros

  • Free with a Google account: 100 notebooks, 50 sources each
  • Answers are grounded in the sources you upload

Cons

  • Usage is compute-based, refreshing every 5 hours up to a weekly limit
  • You can't buy Pro on its own; it's part of the Google AI subscription bundle

Skip it if: You need cross-project memory across many notebooks.

Price: Free tier: yes. Higher limits via Google AI Pro ($19.99/mo). (checked Oct 10, 2026)

Sources for NotebookLM (Gemini Notebook)

3. Innogath

Best for: Researchers wanting cited AI reports.

Pros

  • AI research workspace producing cited reports, visual knowledge maps and notebooks
  • Exports to Markdown, PDF and DOCX with citations

Cons

  • Citations still need checking
  • Pricing not verified

Skip it if: You require verified sources only.

Price: Pricing not verified (checked Oct 14, 2026)

Sources for Innogath

4. Perplexity Spaces (now Projects) Renamed

Best for: Teams running ongoing research topics in Perplexity.

Pros

  • Shared research workspaces with files, context and collaborators
  • Up to 5,000 files and 2.5 GB per project on every plan

Cons

  • Spaces have been renamed Projects, so the listing name is stale
  • Free users get limited uploads

Skip it if: You only do one-off searches.

Price: Included with Perplexity. Collaboration on Pro ($20/mo) and up. (checked Oct 11, 2026)

Sources for Perplexity Spaces (now Projects)

Terminal and agent research

5. rStation

Best for: Researchers and hobbyists who want to point their agents at open math and CS problems.

Pros

  • Free public forums where AI agents work on specific research problems (Riemann hypothesis, no-three-in-line and others) and publish papers
  • Bring your own model and compute: agents join through an API and a skill.md guide, and humans can browse

Cons

  • Reviews are automated; human verification of 'solved' problems is recorded separately
  • Early days: HN commenters noted most posts get 0 to 2 upvotes and asked if it's a 'token burning black hole'

Skip it if: You want curated, human-reviewed research.

Price: Free to browse and join; you supply your own model and compute. (checked Oct 19, 2026)

Sources for rStation

6. Mole

Best for: Developers wanting controlled research agents.

Pros

  • Terminal deep-research agent with an enforced budget and verified quotes
  • Privacy boundary for local data

Cons

  • Young
  • Needs your model API keys

Skip it if: You want a GUI.

Price: Free and open source (Apache-2.0) (checked Oct 14, 2026)

Sources for Mole

Specialized research

7. QED Science

Best for: Academic scientists who want a critical pre-submission review of a paper or grant.

Pros

  • Breaks a manuscript or grant into its core claims and flags gaps in logic, methodology and feasibility, with suggested fixes
  • Free for academic researchers, with an optional insights feed of new findings relevant to your work

Cons

  • Its FAQ says users grant QED a license to optionally train and evaluate its own models on uploads, and learned parameters remain after you delete a project
  • Doesn't detect fraud or data manipulation; it assumes the data are genuine

Skip it if: You can't upload unpublished work to a third party, or you need fraud or image-manipulation checks.

Price: Free for academic researchers; Premium/Validity Machine pricing not published (early access). (checked Oct 13, 2026)

Sources for QED Science

FAQ

What are the best AI Research Assistant Tools in 2026?

Our 7 picks for finding, digesting and citing sources, grouped by task. Research workspaces: NotebookLM (Gemini Notebook), Masala (formerly Panel), Innogath and Perplexity Spaces (now Projects); Terminal and agent research: rStation and Mole; Specialized research: QED Science. Inside each task, tools are ordered by aitrove popularity, not by a quality score.

Which AI tools for AI research assistants have a free tier?

NotebookLM (Gemini Notebook), Masala (formerly Panel), Mole and QED Science have a free tier, according to the sourced pricing on each tool card. For the rest our sources don't state a free plan, so check the vendor page.

How were these AI tools for AI research assistants chosen?

We only publish a job roundup when at least 5 reviewed tools genuinely fit the work. Pros, catches and prices are compiled from public sources (official pricing and docs pages, published reviews, user reports), not hands-on testing, and were checked Oct 14, 2026. Every tool card links its sources.

More roundups

Pros, catches and prices are compiled from public sources (official pricing and docs pages, published reviews, user reports), not hands-on testing. Prices change often, so check the vendor page before you buy.