How is this different from asking a chatbot?

A chatbot answers from memory and cannot show its work. SixSentences_ runs a real search over real indexes, screens each record against criteria that were frozen up front, and logs every decision with a reason. The result is a literature base you can defend, not a plausible paragraph.

What sources does it search?

One search covers the major catalogs at once: more than 240 million works spanning PubMed, Crossref, arXiv and DOAJ, unified through the open OpenAlex index. On top of that it harvests grey literature from the open web and retrieves full texts wherever they are openly available.

What does audit grade mean?

The protocol is written and frozen before the search runs, every include and exclude carries a written reason, and the report shows PRISMA style flow counts that reconcile. You can hand the record to a supervisor, a reviewer or a committee, and every number has a trail.

How does monthly research capacity work?

You always see one percentage for the month instead of a fictional token balance. Usage adapts to the work SixSentences_ actually performs: workflow depth, model choice, sources, figures and the number of records screened or extracted. Higher plans stretch the same percentage further, and the meter resets each month.

Can I get my results out?

Yes, in whatever shape your workflow needs: RIS, BibTeX and CSL JSON downloads, a direct push to Zotero, and a JSON API with webhooks if you want to build on top of it.

We are onboarding researchers in small waves so every workspace gets attention. Leave your email at the bottom of this page and you will get exactly one email when it is your turn.

Systematic literature search, run overnight

Stop searching.Start reading.

One question in. A screened, ranked and fully documented literature base out.

240M+ works · Every decision logged · Ready to cite

Request early access

The name is the method: six sentences from question to literature base.

01/06

Ask

You ask one question in plain language and decide how deep the search should go.

02/06

Protocol

It writes the protocol first. Criteria and search strings are frozen before the first result arrives.

03/06

It searches PubMed, Crossref, arXiv and more in one sweep: 240 million works, plus the open web.

04/06

Screen

It screens every single record against your criteria and writes down the reason for every decision.

05/06

Rank

It ranks what survives by relevance to your question and pulls open full texts into your library.

06/06

Report

You wake up to the record. Flow counts, decisions and citations, ready to defend.

The record

Not a chat transcript. A record you can defend.

SixSentences_ documents the search the way a methods chapter would. The protocol is frozen before the first query runs, every screening decision carries a written reason, and the final numbers reconcile. If a reviewer asks why paper 41 is out, the answer is one click away.

Request early access

app.sixsentences.com

A finished search in SixSentences_: the search string, the PRISMA 2020 flow from 457 records to 30 included, and the ranked works with their screening decisions

A REAL FINISHED SEARCH: 457 FOUND · 420 SCREENED · 30 INCLUDED

01.

The frozen protocol

Inclusion criteria, exclusion criteria and search strings, locked before the search runs. No moving goalposts.

02.

The decision ledger

Every include and every exclude with its reason attached. Not a sample. All of them.

03.

The report

PRISMA style flow counts and citations, exported as RIS, BibTeX, CSL JSON or pushed straight to Zotero.

Between searches

A research assistant that shows its work.

Most days you are not running a review. You are checking a claim, chasing a reference, or reading. SixSentences_ keeps working between the big searches.

01 · grounded

Answers that show sources

The assistant quotes what it cites, cites what exists, and says so when it finds nothing. Every claim links back to a real paper.

02 · library

A living library

Retrieved full texts stay in your workspace. Read them in place, search inside them, and ask questions across everything you collected.

03 · pipeline

API and webhooks

Everything the interface can do, your scripts can do too. Start searches, pull results, and wire the record into your own pipeline.

The workspace

Your questions, your rules.

The morning after a search: your library on the left, the paper you are reading on the right, and a conversation in between that quotes its sources.

app.sixsentences.com

The SixSentences_ workspace: the library and a conversation on the left, the full text of a paper open on the right

01 · model

Choose your model

Every question can run on the model you trust, or on auto, which picks the cheapest one that does the job. You decide which provider sees your text.

02 · tactic

Choose your tactic

A grounded quick answer in seconds, a deep screen over full texts, or the whole systematic method overnight. You set depth, criteria and scope per run.

03 · privacy

Privacy is the house rule

Your protocols, decisions and library stay in your workspace and serve no other purpose. Delete a chat, a search or the whole account any time, and it is gone.

Manuscript Studio

Write the manuscript where the evidence lives.

A structured authoring room with the search wired in: import Word or start from your own template, navigate a live outline, let the assistant propose edits, connect your own sources and research data, and compile the PDF as you write.

app.sixsentences.com

The SixSentences_ Manuscript Studio: an evidence-grounded authoring chat beside the structured source and compiled PDF

THE SEARCH FROM ABOVE, WRITTEN UP: ITS 30 INCLUDES ARE THE BIBLIOGRAPHY

01 · cite

A bibliography with receipts

references.bib combines the include set of your searches with your own PDFs, BibTeX and RIS. Every source receives a real key before the assistant can cite it.

02 · edit

Edits arrive as proposals

Ask for a rewrite and you get a diff: exact find and replace, applied only when you accept it. Highlight a passage in the PDF and discuss exactly that passage.

03 · results

Your data joins the argument

Import CSV, JSON or Excel with provenance and missingness intact. Ask the assistant to use exact values, or turn them into a deterministic chart for the paper.

Sources & exports

Searched in full, exported in yours.

Every search sweeps the major catalogs in one pass: more than 240 million scholarly works spanning PubMed, Crossref, arXiv and DOAJ, unified in one open index. Grey literature joins in from the open web, and full texts are retrieved wherever they are openly available.

Visit openalex.org

OpenAlex

PubMed

Crossref

arXiv

DOAJ

DOI

Zotero

BibTeX

RIS

CSL JSON

JSON API

webhooks

Pricing

Start free. Scale when the reviews do.

One calm meter shows the percentage of monthly research capacity left. Light answers barely move it; deeper searches use more based on models, sources and records actually processed.