From topic to submission, in five steps
ResearchFlow walks every step of the research lifecycle with you. Six bounded AI agents handle the heavy lifting — research, analysis, drafting, editing, and compliance — while keeping your voice, your effort, and your integrity at the center.
Topic
Shape a vague idea into a researchable question
The Topic Agent sits at the very start of your research journey. Drop in a broad field — climate adaptation, social media use, supply chain risk — and it walks you through narrowing it down to a researchable question that is scoped to your timeline, novel enough to matter, and feasible with the resources you have.
- Run scope, novelty, and feasibility checks on every candidate question
- Get a written defense for your topic choice that your supervisor can review
- Lock your question and the agent scaffolds a starter outline into the Writing Editor
- Move from idea to outline in a single sitting
Research
Find, curate, and read sources without leaving Flow
The Sources Curator Agent scans ResearchLib and the open web for relevant papers, pre-prints, theses, and datasets. The Deep Research Agent runs a multi-hop trace on harder questions. The Source Verifier cross-checks every source against its origin. You end this step with a curated, verified library — every entry traceable.
- Sources Curator builds a starter library in minutes — deduped, ranked, metadata-rich
- Deep Research follows citation chains forward and backward until evidence converges
- Source Verifier catches dead links and fabricated citations before you cite them
- Read every source in the in-app e-reader with highlights and annotations
Data
Import, profile, and analyze your datasets
Upload a CSV and the Analysis Agent profiles it — variable types, distributions, missing values. It recommends the right statistical test from the 14 supported, explains why, runs it, and produces a publication-ready chart. You write the interpretation before the result is inserted into your paper.
- 14 statistical tests covering parametric, nonparametric, descriptive, and diagnostic
- Test recommendation based on the actual shape of your data
- Publication-ready charts with proper labels, legends, and effect sizes
- Interpretation required before results are inserted — no black boxes
Writing
Effort-locked drafting that keeps your voice in
The Full Writing Agent expands your outline into structured drafts — but it is effort-locked. Each paragraph is labeled as Student-Authored, AI-Assisted, or Quoted. The Effort Balance Index updates in real time. The Writing Editor handles rich text, inline LaTeX, charts, and images with auto-save and version history.
- Effort-locked — no one-click chapter generation, ever
- Per-paragraph contribution labels feed the Effort Balance Index
- Writing Editor with inline LaTeX, chart builder, and image uploads
- Auto-save with full version history and one-tap restore
Submission
Integrity gates and a one-click export
The Compliance Agent runs four checks — similarity, originality, bias, and citation validation — before you can finalize. If any check fails, finalization is blocked. Once every check passes, generate a downloadable integrity report and export your paper to PDF, DOCX, or LaTeX in your institution template.
- Four integrity checks: similarity, originality, bias, citation validation
- Per-paragraph feedback with suggested fixes inline
- Downloadable, timestamped, reproducible integrity report
- One-click export to PDF, DOCX, or LaTeX with institution-compliant template
Ready to walk the lifecycle yourself?
Sign up free and let ResearchFlow guide you from topic to submission — with integrity built in.