Researchful
ResearchFlow

Your AI-powered research companion

From a vague topic to a submission-ready paper, ResearchFlow walks every step with you. Six bounded AI agents handle research, analysis, drafting, editing, and compliance — while keeping your voice, your effort, and your integrity at the center.

  • Source-bounded AI — no claims without evidence
  • Effort Balance Index per chapter
  • Compliance gates block submission until integrity passes
6 Bounded AI agents
14 Statistical test types
50+ Agentic tools
5 Citation styles
The lifecycle

Five steps from topic to submission

A guided workflow that keeps you moving forward without skipping the hard thinking.

  1. 01

    Topic

    Narrow a broad field into a researchable question with scope, novelty, and feasibility checks.

  2. 02

    Research

    Curate sources from ResearchLib and the open web; cluster themes; extract metadata; fact-check claims.

  3. 03

    Data

    Import datasets, run statistical analyses, generate charts, with interpretation required before insertion.

  4. 04

    Writing

    Effort-locked drafting expands your bullets into paragraphs; contributions are labeled per paragraph.

  5. 05

    Submission

    Compliance gates, integrity report, and a one-click PDF export ready for your supervisor.

The team

Six bounded AI agents, each a specialist

Each agent has a narrow scope, a tool belt of 6 to 20 tools, and guardrails that protect academic integrity.

Research Agent

18 tools
Sources

Source-bounded extraction, theme clustering, and summaries. Searches the web and ResearchLib, then fact-checks every claim against the sources it found.

Analysis Agent

20 tools
Statistics

Recommends statistical tests based on your data shape, runs parametric and nonparametric analyses, and generates publication-ready charts.

Drafting Agent

8 tools
Drafting

Expands your bullet points into coherent paragraphs. Effort locks forbid one-click generation — your voice stays in the writing.

Editing Agent

8 tools
Editing

Grammar, style, citation sync, and TOC generation. Always previews changes before applying, with one-tap revert.

Compliance Agent

8 tools
Integrity

Similarity, originality, bias, and fabricated-citation detection. Blocks finalization until every integrity check passes.

Orchestrator

6 tools
Coordination

Coordinates all agents, enforces gates between phases, and maintains a full audit trail of every action and approval.

Sources Curator Agent

A research librarian that never sleeps

Drop in a topic and the Sources Curator Agent scans ResearchLib and the open web for relevant papers, pre-prints, theses, and datasets. It dedupes, extracts metadata, ranks by relevance and recency, and builds you a starter library in minutes — every entry traceable to its origin.

  • Cross-source deduplication with DOI and title fingerprinting
  • Automatic metadata extraction — authors, venue, year, abstract
  • Relevance ranking with explainable scoring
  • Open-access preference toggle
Explore the Sources Curator
Deep Research Agent

Multi-hop research that follows the trail

Give the Deep Research Agent a single question and it runs a multi-hop trace — breaking the question into sub-queries, reading the results, identifying gaps, and following citation chains until the evidence converges. You get a synthesized brief with every step of the trail visible and reviewable.

  • Multi-hop decomposition with explainable sub-queries
  • Citation chain following — forward and backward
  • Gap detection — surfaces what is missing
  • Reviewable trail — every claim linked to its source
See deep research in action
Full Writing Agent

Drafted with you, not for you

The Full Writing Agent takes your outline, your notes, and your sources and produces a structured draft section by section — but it is effort-locked. Each paragraph is labeled as Student-Authored, AI-Assisted, or Quoted, and the Effort Balance Index updates in real time so you always know how much of the chapter is genuinely yours.

  • Section-by-section drafting from your outline
  • Effort-locked — no one-click chapter generation
  • Per-paragraph contribution labels
  • Live Effort Balance Index per chapter
Understand the Writing Agent
Source Library

An in-app e-reader for every source you collect

Read, highlight, annotate, and cite — without ever leaving ResearchFlow.

In-app e-reader

Read PDFs and web sources in a built-in reader with highlight, note, and bookmark tools. Annotations sync to your outline automatically.

Citation export

Export any source to APA, MLA, Chicago, Harvard, or IEEE in one tap. Citation metadata is verified before it lands in your bibliography.

Source picker

When the Drafting Agent proposes a claim, it surfaces candidate sources from your library. You pick the one that fits — no citation is added without your approval.

Writing Editor

A writing surface built for academic work

Rich text, inline LaTeX, a chart builder, and image uploads — all in one editor that auto-saves every keystroke and keeps a full version history. The toolbar stays out of your way until you need it, and every paragraph can be tagged with its contribution label.

  • Rich text with headings, lists, tables, and callouts
  • Inline LaTeX rendering for equations and symbols
  • Built-in chart builder with 12 chart types
  • Image upload with caption and figure numbering
Explore the Writing Editor
Data Analysis

14 statistical tests, one guided workflow

Upload a CSV, let the Analysis Agent recommend the right test, and produce a publication-ready chart. You write the interpretation before the result is inserted.

Supported tests
Independent t-testPaired t-testOne-sample t-testOne-way ANOVATwo-way ANOVARepeated ANOVAChi-square goodnessChi-square independencePearson correlationSpearman correlationSimple regressionMultiple regressionMann-Whitney UKruskal-Wallis
Collaboration

Bring your supervisor into the draft

Share a project with your supervisor with read or comment permissions. They can leave inline annotations, resolve threads, and join a live meeting mode where you walk through the draft together in real time. Every comment is timestamped and tied to a paragraph — no more email chains with confusing attachment versions.

  • Share with supervisor — read or comment permissions
  • Live meeting mode with shared cursor and voice
  • Inline annotations tied to paragraphs
  • Resolved-thread tracking with audit history
See collaboration features
Compliance

Integrity gates that protect your defense

The Compliance Agent runs four checks before you can finalize. If any fails, you fix it first.

Similarity check

Compares every paragraph against public sources and your own previous work. Reports a per-paragraph similarity score with the matched text highlighted.

14% — within threshold

Originality

Detects patterns consistent with AI-only generation — flat perplexity, repetitive structure, and absence of personal voice. Flags paragraphs that need rewriting.

8% — within threshold

Bias detection

Scans language for gendered, cultural, or leading-word bias. Suggests neutral alternatives inline and explains why a phrase may be read as biased.

3% — within threshold

Citation validation

Verifies every in-text citation against a real source and flags fabricated references. Cross-checks the bibliography against the body — both directions.

All citations verified
Pricing

Pricing that grows with your research

Start free. Upgrade when your project grows. Cancel anytime.

Free

For students starting their first research project.

$0 / forever
Get Started Free
  • 1 active project
  • Research Agent — 25 runs/mo
  • Source Library — 50 sources
  • APA + MLA citation styles
  • PDF export with watermark
  • Community support

Researcher

For postgraduates, supervisors, and research groups.

$19.99 / per month
Start 14-Day Trial
  • Unlimited projects
  • All 6 AI agents — unlimited runs
  • Source Library — unlimited
  • All 5 citation styles + custom
  • PDF + DOCX + LaTeX export
  • Data Analysis — unlimited
  • Collaboration — 10 reviewers
  • Live meeting mode + audit trail export
  • Dedicated support
FAQ

Questions about ResearchFlow

Real answers to the questions students ask before they sign up.

Is the writing really mine, or does the AI write the whole thing?

The Drafting Agent is effort-locked. It expands your bullet points into paragraphs, but it never writes a chapter from a single prompt. An Effort Balance Index shows your contribution vs AI assistance per chapter, and every paragraph carries a label: Student-Authored, AI-Assisted, or Quoted.

How does Flow handle citations and references?

The Editing Agent syncs in-text citations to your bibliography in APA, MLA, Chicago, Harvard, or IEEE. The Compliance Agent also validates every citation against real sources and flags fabricated references before you can finalize.

Can my supervisor see what I am working on?

Yes. You can share a project with your supervisor with read or comment permissions. They can leave inline annotations, resolve threads, and view the audit trail of agent actions. A live meeting mode supports real-time co-presence during reviews.

What kinds of statistical analyses are supported?

The Analysis Agent supports 14 test types including t-tests, ANOVA, chi-square, regression, Mann-Whitney, Kruskal-Wallis, and more. It recommends the right test based on your data shape, runs it, and produces a chart. You must write the interpretation before the result is inserted into your paper.

Do I need to install anything to use Flow?

No. Flow runs entirely in your browser, mobile-native first. It works offline with queued operations that sync when you reconnect. There is nothing to download, install, or update.

What happens to my data and my sources?

Your data is yours. Every source you upload, every note you take, and every draft you write is stored against your account with full export rights. We never train models on your private content, and you can request a full data export at any time.

Start your research project today

Sign up free, define your topic, and let ResearchFlow walk you through every step — with your effort, your voice, and your integrity intact.