Skip to content
neuroflow0.2.31 · docs

← neuroflow

/neuroflow:grant-proposal

Write a competitive, funder-ready grant application — from research question to full draft.

/grant-proposal is the neuroflow command for writing complete grant applications. It automatically discovers your ideation outputs, fetches funder call documents from URLs, adapts its structure to the target funder, and drafts every section with word-count tracking and review-criteria alignment.

When to use it

  • You need to write a funding application for your neuroscience study
  • You want to turn an ideation output into a full grant application
  • You have a funder call URL or PDF and want Claude to parse the requirements
  • You need to draft specific sections (Specific Aims, Approach, Budget) with expert guidance
  • You want your draft checked against the funder's actual review criteria

What it does — step by step

Step 0: Discovers your research idea automatically

Before asking you anything, Claude checks .neuroflow/ideation/ for existing research question documents. If it finds them:

I found your ideation output. Here's what I'll build the grant from:
• Research question: Does white noise reduce P300 amplitude in a visual oddball paradigm?
• Modality: EEG (64-channel, 1000 Hz)
• Population: Healthy adults, 18–35
• Preliminary data: Pilot N=8, effect size d=0.61 (SE=0.18)
• Key references: Polich (2007), Näätänen et al. (2018)

Is this the right idea to build the grant around?

If no ideation files exist, Claude asks you for the research question and offers to read a funder call URL.

Step 1: Fetches and parses the funder call

Provide a URL to the funding call, or paste the call text. Claude extracts:

  • Scheme name, budget ceiling, duration
  • Page/word limits per section
  • Required sections and order
  • Review criteria
  • Eligibility constraints
  • Deadline
  • Submission portal

Then confirms everything with you before drafting a single word.

Step 2: Builds a proposal outline

Claude produces a structured outline adapted to the funder's requirements — with page budgets next to each section — and asks for your approval before proceeding.

Step 3: Drafts section by section

Claude drafts one section at a time, showing: - Which review criteria the section addresses - Word/page count vs. the limit - A menu of options after each section: revise / expand / next / save / checklist

Step 4: Quality check

Before saving, Claude runs a quality checklist: - Every aim is measurable - Power analysis is present - Limitations and alternatives are addressed - All sections are within word/page limits (word counts come from a tool, not an estimate) - Review criteria are explicitly addressed in the text - References come from your project library, and their DOIs resolve (the cite_check.py script, which also looks for retraction notices in Crossref)

If the funder asks applicants to declare generative-AI use, Claude drafts the declaration from the session logs (what the model did, what you did) and you confirm it. The humanizer is a style edit that runs only when you ask for it — never a way to hide AI involvement.

Step 5: Saves to memory

The completed draft is saved as grant-[funder]-[date].md in .neuroflow/grant-proposal/. With your confirmation, the deadline goes to timeline.md and funder and scheme are noted in project_config.md.

Supported funders

Funder Scheme Typical budget Duration
NIH R01 $500K DC/yr 5 years
NIH R21 $275K total 2 years
ERC Starting Grant €1.5M 5 years
ERC Consolidator Grant €2M 5 years
ERC Advanced Grant €2.5M 5 years
Wellcome Trust Discovery Award £3–5M 5 years
Horizon Europe EIC Pathfinder €3M 4 years
National research council or science foundation Standard project grant Varies by call Often 3 years

Any funder works

If your funder is not in the table above, Claude will ask you for the requirements directly, or read a URL or PDF you provide. For national funders it checks the call text for language and format rules — some parts may have to be in the national language and others in English.

Neuroscience-specific expertise

/grant-proposal understands neuroscience methodology. When drafting the Approach section it asks for and includes:

Modality Detail included
EEG / ERP Cap system, sampling rate, bandpass, epoch, ICA, software
fMRI / BOLD TR, field strength, HRF, motion scrubbing, cluster correction
iEEG / SEEG Electrode type, HFO analysis, reference scheme
Eye-tracking Tracker model, fixation algorithm, AOI definition
ECG / HRV R-peak detection, SDNN/RMSSD, artifact correction
Behaviour Timing precision, counterbalancing, trial counts

A formal power analysis is always included — using your pilot data or published effect sizes.

Example session

/neuroflow:grant-proposal

Claude: I found your ideation output. Here's what I'll build the grant from:
        • Research question: Does white noise reduce P300 amplitude in a visual oddball task?
        • Modality: EEG (64-channel)
        • Population: Healthy adults 18–35, N=40
        • Preliminary data: Pilot N=8, d=0.61
        • Key refs: Polich (2007), Näätänen et al. (2018)

        Is this the right idea to build the grant around?

You: Yes. Applying to my national research council's standard project grant.

Claude: Do you have a URL to the current call for proposals?

You: https://research-council.example.org/calls/2026-standard/

Claude: [reads URL]
        Standard Project Grant 2026:
        • Budget: up to €200K/year, 3 years
        • Sections: Project summary (600 words, national language + English),
          State of the art, Objectives, Methodology, Feasibility,
          Budget justification, Timeline
        • Review criteria: Scientific originality, Feasibility, Team qualifications, Budget
        • Deadline: October 15, 2026
        • Submission: the funder's online portal

        Confirmed — shall I build the outline?

You: Yes

Claude: [produces outline with page budgets]

        Ready to draft Section 1: Project Summary (600 words)?

You: Go

Claude: [drafts Project Summary]

        Word count: 587 / 600  ✅

        Options: revise | expand [topic] | next | save | checklist

Output files

File Location Description
Grant draft .neuroflow/grant-proposal/grant-[funder]-[date].md Full application
Section drafts .neuroflow/grant-proposal/draft-[funder]-[date]-[section].md Per-section files during drafting
AI-use declaration .neuroflow/grant-proposal/ai-use-statement-[funder]-[date].md Drafted from the session logs, confirmed by you
Flow index .neuroflow/grant-proposal/flow.md Funder, scheme, deadline, section status
Session log .neuroflow/sessions/YYYY-MM-DD.md Appended automatically

Files read and written

Direction Files
Reads .neuroflow/project_config.md, .neuroflow/flow.md, .neuroflow/objectives.md, .neuroflow/ideation/, .neuroflow/grant-proposal/flow.md, .neuroflow/sessions/ (AI-use declaration)
Writes .neuroflow/grant-proposal/, .neuroflow/grant-proposal/flow.md, .neuroflow/objectives.md, .neuroflow/reasoning/grant-proposal.jsonl, .neuroflow/timeline.md, .neuroflow/sessions/YYYY-MM-DD.md
  • /ideation — define your research question first if you don't have one
  • /experiment — design the paradigm that feeds into the Approach section
  • /preregistration — register your analysis plan; reference it in the grant
  • /write-report — generate progress reports for funders after funding is secured
  • /finance — manage the awarded budget after the grant is funded