Consensus is famous for its positioning: ask one question, see what the research says. This post compares Consensus with Phở Chat using public data (first published 11 July 2026, prices rechecked 29 October 2026), including the places where we think Consensus does well.
What Consensus is and where it shines
Consensus is a scientific question answering tool: you ask a question, the system finds and synthesises the relevant research. Their search engineering is quite serious, following a three-step pipeline the company has published: combined semantic and BM25 search, re-ranking around 1,500 papers by research quality signals, then selecting the 20 most precise. The signature Consensus Meter summarises the level of agreement (yes, no, mixed) across studies for yes-no questions, and Deep Search lets an agent run up to 20 searches and review more than 1,000 papers for a single query.
On scale: at its $11.5 million Series A (July 2024, led by Union Square Ventures), the company reported more than 400 thousand monthly active users. The corpus of around 200 million documents includes preprints; not all of it is peer-reviewed work.
Quick comparison table
| Criterion | Consensus | Phở Chat |
|---|---|---|
| Product focus | Scientific Q&A + consensus level | Research and medical workspace with verified citations |
| Answer style | Synthesis of top 20 papers, Consensus Meter | Answers ranked by evidence tier, verified sentence by sentence |
| Corpus | ~200M documents (including preprints) | PubMed, OpenAlex, arXiv, Europe PMC in parallel |
| Deep workflows | Deep Search (up to 20 searches) | AI title/abstract screening on Pro (60 records a run, a reason for every verdict) + PRISMA counts, RoB2/GRADE appraisal, meta-analysis |
| Writing and export | Synthesis reports | Structured answers with citations, BibTeX/RIS export |
| Bring your own API key (BYOK) | No | Yes: OpenAI, Anthropic, Google, DeepSeek, xAI, Groq or OpenRouter keys, zero markup |
| Vietnamese interface | No | Yes, one of 14 interface languages, medical terminology included |
| Price | Free plan; Pro $20/month or $144/year; Deep $65/month | Free $0 · Pro $10/month or $99/year (your own model key) |
| Payment | International cards | International cards + VietQR in Vietnam |
When to choose Consensus
You need to know quickly “what the research says” about an English yes-or-no question, for example whether an intervention is effective. The Consensus Meter gives a very fast overview. You mainly need the initial survey step, not screening or deep appraisal yet.
When Phở Chat is the better fit
You need to know exactly which source backs each sentence in the answer: Phở’s cite-or-abstain mechanism labels any sentence without a real source as “unverified” instead of letting the AI fabricate one. You need to go further after the Q&A step: systematic review screening, methods appraisal, meta-analysis, the steps Consensus does not do. You want cost control: bring your own OpenAI, Anthropic, Google, DeepSeek, xAI, Groq or OpenRouter API key (BYOK, zero markup); Pro is $99 a year on top of what you pay the provider. And if you work in Vietnamese, Phở answers natively with correct medical terminology.
An honest note: Consensus figures come from the company’s official blog (search pipeline, Deep Search, user numbers at the July 2024 Series A) as of 11 July 2026. User numbers and benchmarks are self-reported by the vendor. We update this page periodically.
Updated 4 November 2026: competitor prices now come from our 29 October check of each vendor’s own pricing page; Phở’s rows describe what is live today. Other competitor facts are as of 11 July 2026 unless a line says otherwise.