Consensus vs Elicit
Both search academic papers, but the workflows differ. Consensus gives quick evidence synthesis with a Consensus Meter showing where retrieved studies agree; the display is not proof the evidence is complete or the claim is true. Its Free tier costs $0 and includes unlimited Papers searches, 10 Pro messages and 3 Deep reviews per month; Pro is $20/month (or $144/year billed annually, equivalent to $12/month) with unlimited Pro messages and 15 Deep reviews per month; Deep is $65/month (or $540/year billed annually, equivalent to $45/month) with 200 Deep reviews per month. Elicit builds structured extraction tables across papers. Its Basic tier is free with limited Research Agent and Research Report usage but unlimited search, summaries and chat with full-text access; the Dedicated Systematic Review Workflow requires Pro, Scale or Enterprise. Pro (Industry) is $75/user/month billed monthly, or $49/user/month billed as $588 annually; Scale (Industry) is $375/user/month billed monthly, or $229/user/month billed as $2,748 annually; separate Academic and nonprofit pricing also exists (see the Elicit support article); Enterprise is custom. Currency is not explicitly labeled on either vendor's pages. Quick consensus check: Consensus. Screening and data extraction for a review: Elicit. No measured accuracy winner.
| Decision point | Consensus | Elicit |
|---|---|---|
| Best suited to | People who need a first-pass map of published research with citations to verify. | Researchers who want to structure academic searches, extract study details into a table, or export references for checking. |
| Plans and cost | Free tier: $0. Pro: $20/month or $144/year billed annually (equivalent to $12/month). Deep: $65/month or $540/year billed annually (equivalent to $45/month). Currency is not explicitly labeled. Student, faculty and clinician discounts up to 40% for eligible users. | Basic is free with limited Research Agent and Research Reports usage. Pro (Industry): $75/user/month billed monthly, or $49/user/month billed as $588 annually. Scale (Industry): $375/user/month billed monthly, or $229/user/month billed as $2,748 annually. Enterprise is custom. Separate Academic and nonprofit pricing exists; see Elicit support. Currency is not explicitly labeled; confirm the applicable plan and usage limits at checkout. |
| Limitations | The database is not comprehensive, AI can misread papers, and full-text access depends on the source. Full-text AI analysis is available for some papers through publisher partnerships, but opening the article itself may still require an institutional or personal subscription. | Usage limits and export options depend on plan, and evidence and full-text availability vary by paper. The named Systematic Review workflow requires a Pro, Scale or Enterprise plan; Basic covers Research Agent sessions and Research Reports with limited usage. Review extracted values and citations against the original paper before relying on them. |
Choose by the job
- You need a fast answer to a focused question about what published studies say, with papers you can open: start with Consensus. Search the paper database and read the cited studies; the Consensus Meter shows where retrieved studies agree, but agreement display is not proof the evidence is complete or the claim is true.
- You need a structured extraction table across papers for a review, with screening decisions you can defend: start with Elicit. Use the named Systematic Review workflow on Pro, Scale or Enterprise to screen papers, extract values into columns, and export the table; Basic does not include that workflow.
- You need a quick overview of a topic space before deciding whether a full review is worth doing: start with Elicit. Run a Research Report on any plan, including Basic with limited usage, then check the cited papers before treating the overview as findings.
- You need to prove that no relevant study exists, or to cover every database for a formal systematic review: start with Consensus. Neither tool alone settles this: Consensus states its database is not comprehensive, so absence from results does not prove no study exists; a formal review needs a documented protocol across sources.
Trade-offs that change the choice
| Check | Consensus | Elicit |
|---|---|---|
| Evidence domain | A research-paper database (220M+ papers per product docs) with cited synthesis and an agreement display; it does not search the broader web. | Academic papers plus uploaded resources, organized into tables; the Research Agent can also reach the larger web, but extracted claims still need checking against the original paper. |
| Free and paid thresholds | Free: unlimited Papers searches, 10 Pro messages, 3 Deep reviews, 10 Study Snapshots and 30 API & MCP calls per month. Pro: $20/month or $144/year billed annually (equivalent to $12/month) with unlimited Pro messages and 15 Deep reviews; Deep: $65/month or $540/year billed annually (equivalent to $45/month) with 200 Deep reviews. Currency is not explicitly labeled. | Basic: free with limited Research Agent and Research Reports usage. Pro (Industry): $75/user/month billed monthly, or $49/user/month billed as $588 annually. Scale (Industry): $375/user/month billed monthly, or $229/user/month billed as $2,748 annually. Enterprise: custom. Separate Academic and nonprofit pricing exists; see the Elicit support article. Currency is not explicitly labeled. |
| Full-text and source reach | AI analysis can use full text via publisher partnerships, but opening the article itself may still need an institutional or personal subscription; where full text is unavailable, results fall back to metadata, abstract or indexed information. | Search covers 138M+ papers with full-text chat where available; finding a paper does not mean all contents are available, and extracted values must be checked against the source text. |
| Extraction and export Sources: Elicit exports and plan limits ↗ | Cited synthesis with supporting passages and an agreement display; export is not the product's structured-review path — use the cited papers as the handoff to a review table. | Reports export as PDF or Word on any plan; tables export as CSV or Excel; references as RIS or BIB. Systematic Review tables need Pro, Scale or Enterprise; the export guide names 'Plus and higher' for agent-session tables, a plan name absent from the Industry pricing view captured on 2026-10-26 (an earlier 2026-10-21 capture listed a Plus plan). |
| Review steps | Define the question, search papers, open cited sources, check design, population, results and limitations, and record what the retrieved set does not cover. | Define the question and inclusion/exclusion criteria, choose the workflow, screen papers, extract values into columns, check each value against source text, then export the table and references. |
| Unconfirmed items | Paid checkout currency is not labeled on the plan pages; the pricing-page hero states 400M+ papers while product docs state 220M+ — the two figures are not reconciled here. | Monthly and annual Industry views show different per-month figures (Pro (Industry) $75/user/month billed monthly vs $49 monthly equivalent; Scale (Industry) $375/user/month billed monthly vs $229 monthly equivalent); the annual figures are equivalents, not monthly charges. Which current plan covers agent-session table export is unconfirmed because the export guide names 'Plus', absent from the Industry pricing view captured on 2026-10-26. |
Run your own like-for-like check
This is a proposed evaluation protocol. No new head-to-head product run is reported here.
- Write the question, population or topic, outcome, and the claims that would change the conclusion. Decide whether the job is a fast synthesis or a structured review table.
- Search Consensus for the paper set, or start Elicit's Research Agent or Research Report for an overview; for a controlled review, choose Elicit's Systematic Review workflow on Pro, Scale or Enterprise.
- Screen the retrieved papers against inclusion/exclusion criteria and record the decisions; do not treat a single synthesis as the screening.
- Open each relied-on paper. Check design, population, results, limitations and full-text availability, and record the exact passage behind each extracted value.
- Build the deliverable: a cited synthesis with its coverage limits, or an extraction table with references exported as CSV, Excel, RIS or BIB.
- If the evidence is thin or contradictory, narrow the conclusion to what the retrieved papers support and list the unanswered questions; do not fill gaps with a generated summary.
Record these outcomes
- The question, inclusion criteria and evidence type are written before searching.
- Every extracted value has an opened original paper behind it, not just a citation.
- Plan limits are recorded: Consensus Free 3 Deep reviews per month; Elicit Basic excludes the Systematic Review workflow.
- Full-text gaps are marked; abstract-only evidence is not presented as full-text review.
- The deliverable is exported (table as CSV or Excel, references as RIS or BIB) or the synthesis states its coverage limits.
- Unconfirmed items (checkout currency, Elicit agent-session table export plan, corpus-size conflict) are listed as unresolved, not guessed.
- The method is documented well enough to reproduce: query, date, sources, screening decisions. See the research task pack for the worked ledger format.
Take the comparison into a workflow
Sources for these trade-offs
- Consensus Free, Pro, Deep, team, and API/MCP plan limits ↗ · reviewed 2026-10-26
- Consensus search and cited research synthesis description ↗ · reviewed 2026-10-26
- Consensus paper coverage and variable full-text availability ↗ · reviewed 2026-10-26
- Elicit plans, monthly and annual billing views ↗ · reviewed 2026-10-26
- Elicit workflow selection and Systematic Review plan gate ↗ · reviewed 2026-10-26
- Elicit exports and plan limits ↗ · reviewed 2026-10-26
- Elicit Academic and nonprofit pricing ↗ · reviewed 2026-10-26
Use the same input
Ask 'Is coffee linked to lower depression risk?' Compare how fast each shows the direction of evidence and whether you can export a comparison table.
What a successful result must preserve
Results must come from peer-reviewed papers, show where retrieved studies agree without overstating that display, and export to a review workflow.
Evidence and test status
Pricing and feature comparison based on the vendors' official pricing and help pages (Consensus subscription plans; Elicit pricing and workflow guides), captured on 2026-10-26. Third-party pricing reads were not used as a current-price basis. We did not run a scored hands-on test; no quality ranking is claimed.
- Consensus — Consensus official page ↗
- Elicit — Elicit plans and usage ↗ (reviewed 2026-10-26), Elicit research workflow selection ↗ (reviewed 2026-10-26), Elicit exports and plan limits ↗ (reviewed 2026-10-26), Elicit workflow changes and legacy sessions ↗ (reviewed 2026-10-26)