Skip to content

Side-by-side comparison ยท Research & Search

Elicit vs Exa vs Kagi

Everything that matters when choosing between them, in one place: our rating, what each can do, what they cost, and the trade-offs.

Elicit

3.8

From $49/user/mo

Exa

4.0

From From $4 per 1

Best overall

Kagi

4.3

From $5/mo

Comparing 3 tools, the maximum.

At a glance

Pick it if: Researchers, students and analysts who need to review a lot of papers carefully.

Exa4.0
Best value

Pick it if: People using Claude Code or building their own assistants who want real web search behind them.

Kagi4.3
Best overallEasiest to use

Pick it if: People who search all day for work and are tired of ads and spammy results.

Kagi scores highest overall, but Exa gives you more for your money. Which one is right comes down to what you value most.

Overview

AttributeElicitExaKagi
In one lineFinds, screens and summarises research papers into a comparison table.A search engine built for AI tools, so Claude can find exactly the pages it needs.A paid search engine with no ads or tracking, plus an AI assistant that cites sources.
Pricing modelFreemiumFreemiumPaid
Starting price$49/user/moFrom $4 per 1$5/mo
Free plan Yes Yes No
Works with Claude Code No Yes Yes

Rating breakdown

Our scores out of 5. The higher bar in each row is highlighted.

AttributeElicitExaKagi
Overall rating3.84.04.3Winner
Ease of useHow quickly someone non-technical gets value from it.
3.5
3.0
4.5
FeaturesHow completely it covers its own job, and how well the pieces fit together.
4.5
4.0
4.5
Output qualityHow good the results are before you fix them.
4.5
4.0
4.5
Value for moneyWhat you get for the price, including the free plan.
3.0
4.5
3.5
Speed & reliabilityHow fast it is and how often it falls over.
3.5
4.5
4.5

Feature comparison

A tick means it does this today. A dash means we haven't checked.

AttributeElicitExaKagi
Access & platform
Free plan Yes Yes No
Mobile app No No Yes
Desktop app No No No
Browser extension No No Yes
API access Yes Yes Yes
Teams & business
Team workspace Yes Yes Yes
Enterprise plan or SSO Yes Yes No
Connects to other apps Yes Yes Yes
Open source or self-hosted No No No
AI abilities
Works with your files Yes No Yes
Understands imagesโ€“ No Yes
Makes images No No No
Makes video No No No
Writes code No No Yes
Runs agents or automations Yes Yes No
Learns your brand or style No Noโ€“
Works in many languagesโ€“ Yes Yes
Capabilities score7/ 177/ 179/ 17

Pricing

Plans as listed on each vendor's site when we last checked. Confirm before you buy.

AttributeElicitExaKagi
Plans

Basic

Free

Searching and reading a few papers.

FreeOur pick

Free

Personal use inside Claude Code.

Trial

Free

Seeing the difference.

ProOur pick

$49/user/mo billed yearly

Anyone doing a real literature review.

Pay as you go

From $4 per 1,000 searches

Anyone whose usage outgrows the free credit.

Starter

$5/mo

Light searchers.

Scale

$169/user/mo billed yearly

Teams running many reviews at once.

Enterprise

Custom

Companies building products on top of it.

ProfessionalOur pick

$10/mo

Anyone who searches every day.

Enterprise

Custom

Universities, pharma and large research groups.

Ultimate

$25/mo

Heavy AI users.

Duo and Family

Duo $14/mo, Family $20/mo

Couples and households.

Which plan to pickStay on Basic unless you have a review to run. When you do, Pro has everything one person needs, and the yearly commitment is the only way to get the quoted price. Scale is for teams, not for a solo researcher who wants more headroom.Start free. The monthly $10 of credit is plenty for one person using it from Claude Code, and if you do run out you only pay for what you use. Enterprise is for software companies, not for a business that just wants better search.Use the free trial for a week, then take Professional. 300 searches on Starter goes in a couple of weeks for most people. Only pay for Ultimate if you'd otherwise pay for an AI chat subscription, since it bundles the top models. Yearly billing saves 10%.

Key features

AttributeElicitExaKagi
Standout features
  • Research Reports
  • Extraction tables
  • Systematic Review workflow
  • Chat with papers
  • Routines
  • Library and sharing
  • Meaning-based search
  • Page contents
  • Find similar
  • Deep search
  • Agent runs
  • Hosted MCP
  • Ad-free results
  • Personalised ranking
  • Lenses
  • Kagi Assistant
  • Summaries
  • API and MCP

Pros and cons

AttributeElicitExaKagi
Pros
  • Extraction tables save days on a literature review
  • Every answer links back to the passage in the paper
  • Systematic review workflow records why papers were included or excluded
  • Routines keep a review current without you remembering to check
  • Free plan is useful for reading and summarising
  • Finds pages by meaning, so vague descriptions still work
  • Free credit resets every month and needs no card
  • Hosted connection for Claude Code takes minutes to set up
  • Returns clean page text ready for an AI to read
  • Pay as you go with no subscription
  • No ads, sponsored results or tracking
  • You control which sites show up and where
  • AI assistant with a choice of top models
  • Fair pricing: unused months are credited back
Cons
  • $49 a user a month is a lot for anyone outside research
  • Only academic papers; no business, news or grey literature (reports published outside journals)
  • Extractions still need spot-checking against the source
  • Reports and large screens take a while to run
  • Collaboration is locked to the $169 Scale plan
  • Not much to use on its own; you need Claude or another assistant in front of it
  • Setup assumes some comfort with plugins or settings files
  • Per-request pricing is hard to predict if an agent runs away with searches
  • No mobile or desktop app, because it is a service rather than a product
  • It costs money, and Starter's 300 searches run out fast
  • Top AI models only on the $25 Ultimate plan
  • Local results like maps and shops are weaker than Google's
  • API and MCP use is billed separately from the search plans

Who each one is for

AttributeElicitExaKagi
Best for
  • Literature reviews
  • Extracting data from papers
  • Screening papers by criteria
  • Comparing studies in a table
  • Web search inside Claude Code
  • Finding sites like one you know
  • Building lists of companies or articles
  • Pulling page text for research
  • Everyday search without ads or tracking
  • Research where you want trustworthy sources
  • Hiding spammy or low-quality sites for good
  • Trying several AI models in one assistant
You'll find it used by
  • Researchers
  • Analysts
  • Clinicians
  • Postgraduate students
  • Developers
  • Founders
  • Researchers
  • Claude Code users
  • Researchers
  • Developers
  • Writers
  • Privacy-minded professionals

Questions people ask

Which is better, Elicit, Exa or Kagi?

Kagi scores highest overall in our ratings (Elicit 3.8/5, Exa 4.0/5 and Kagi 4.3/5). Elicit is best for literature reviews and extracting data from papers. Exa is best for web search inside Claude Code and finding sites like one you know. Kagi is best for everyday search without ads or tracking and research where you want trustworthy sources.

Which is cheaper, Elicit, Exa or Kagi?

Elicit: There's a free plan, and paid plans start at $49/user/mo. Exa: There's a free plan, and paid plans start at From $4 per 1. Kagi: There's no free plan. Paid plans start at $5/mo.

Which is easier to use, Elicit, Exa or Kagi?

Kagi rates highest for ease of use in our scores. People who search all day for work and are tired of ads and spammy results.

How we compare

Ratings are our own, from using each tool, across ease of use, features, output quality, value for money, speed & reliability.

Swap in another tool

More tools from Research & Search in the library.