AI assistants and search decision
Consensus
A basic, narrow research-search + summarization workflow is feasible for a single developer in about a week, but the full commercial product (curated corpus, scale, and enterprise features) is not reasonably replicated without more resources.
Visit website↗Open-source builds that already do this
Every project below is open source and already does this job today. Fork one, self-host it, or take the parts you need - the build prompt further down assumes an empty file, and this is the shortcut past that. Licences differ; check the one on each card before you ship. All Consensus alternatives, with the arithmetic →
What a replacement has to do
- Accept a research query → retrieve matching papers from public APIs → rank and filter results → generate concise AI summaries and extract study metadata → present sources with links and exportable citations.
What it still won’t have
- Proprietary curated corpus and any paywalled/full-text indexing
- Enterprise/team features (SSO, org billing, admin controls)
- Any undisclosed proprietary ranking or scoring model
- Scale, reliability, and polished UX of the commercial product
What remains hard
- Product polish and ongoing maintenance
First-year cost
No published price
Consensus does not publish a price we could read, so there is nothing to compare against. What building costs is below.
Money you would actually spend
Time you would spend
—
What you would spend
What we assumed
The verdict above measures whether you could build it. This one is only about money.
Runnable build prompt
Build a minimal AI research assistant using: React frontend, FastAPI backend, Postgres, Redis, and a worker (RQ/Celery). Core features in scope: accept text queries; retrieve paper metadata from public APIs (OpenAlex and arXiv); normalize and store results in Postgres; rank/filter by date and basic study-design tags; call an LLM API to generate 2–3 sentence evidence-backed summaries and extract DOI, sample size, and study type; UI to list results with summaries and source links and export citations as BibTeX/CSV. Out of scope: paywalled full-text scraping, enterprise SSO/billing, and large-scale distributed indexing. Include input validation, error handling for API failures, retry/backoff for background fetches, logging, and automated tests for the retrieval and summarization endpoints.
How we checked
How the score was reached
- Partly verdict base52
- An open-source build was found+5
- 4 cited sources+3
- 3/3 assessment runs agreed+4
- Evidence score64
The base comes from the verdict. Everything under it is a check that either happened or did not, and each one is a fact frozen in this record rather than a judgement made at render time - so the same evidence always produces the same number.
How scoring works →Cited sources · 4
Every page the run actually retrieved.
- official productConsensus: AI for Research
- official pricingConsensus Pricing
- open sourcedzhng/deep-research
- open sourcearc53/DocsGPT
Integrity checks
What held up, and what did not.
