A local-first VS Code extension

Turn a study idea into a reproducible experiment.

CogniSpec helps researchers and students design, validate, and export cognitive experiments from Cursor or VS Code, without hiding the choices that make a study reproducible.

Install CogniSpec See the workflow
working-memory-pilotcognispec-study.json

N-back · working memory

Versioned study specification
Validated
Trials80
Response window1,500 ms
Target rate30%
Random seed48102
✓Technical constraints pass. Ready to generate.
Local firstStudy specifications and participant data stay on your machine.
Reproducible by defaultSave the specification and random seed alongside every study.
Export readyGenerate a runnable task, versioned config, README, and local CSV export.
A visible workflow

From a structured idea to a study you can inspect.

CogniSpec keeps the parts that matter for technical implementation explicit, versioned, and close to your code.

1

Choose a task family

Start with N-back, Stroop, Flanker, or Go/No-Go and configure the task rather than prompting for a generic experiment.

2

Set the constraints

Make timing, trials, balance, response rules, and your random seed visible in the study specification.

3

Validate before generating

Catch technical constraint problems before a runnable task is created, including timing and required export fields.

4

Generate, pilot, revise

Open the self-contained experiment, pilot it, and preserve the configuration that produced it.

Built for the technical work

More structure than an experiment-shaped webpage.

General tools can produce code. CogniSpec focuses on the technical details researchers need to review, preserve, and revise.

◫

Versioned specifications

Keep the study configuration, random seed, and generated materials together in your project folder.

✓

Technical validation

Check trial counts, timing, response windows, balance bounds, seeds, and expected CSV fields before generation.

↗

Runnable jsPsych output

Generate a self-contained index.html task with a readable README and study configuration.

⌁

Analysis-ready local CSV

Participant-level CSV data downloads locally in the browser at the end of a completed task.

⌘

Inside your editor

Use the command palette from VS Code or Cursor, alongside the rest of your research project.

◌

No telemetry

CogniSpec does not collect telemetry or transmit study specifications or participant data.

V0.1 task families

Established starting points, with the rules in view.

Built-in templates are technical implementation starters. Your methods, piloting, accessibility, and interpretation remain your responsibility.

CogniSpec helps implement an experiment. It does not establish construct validity or replace ethics review.

ParadigmCognitive domainGenerated response rule
N-backWorking memoryF match / J non-match
StroopSelective attentionRBGY for ink color
FlankerInhibitory controlF left / J right center arrow
Go/No-GoResponse inhibitionF for green go signals; withhold for red no-go signals
Research integrity

Useful boundaries are part of the product.

CogniSpec supports technical implementation. It deliberately does not turn a passing validation result into a claim about scientific validity.

✓
What CogniSpec validatesTechnical constraints such as trial counts, timing, response windows, task balance, numeric seeds, and expected CSV fields.
↗
What you still need to doChoose a paradigm, pilot it, review accessibility, obtain consent and ethics approval when needed, recruit appropriately, and interpret results responsibly.
⌂
Where your data staysCogniSpec writes files only to the folder you select. Study specifications and participant data are not transmitted by the extension.
FAQ

Questions worth answering directly.

CogniSpec is intentionally focused. Here is what it does, and what it does not claim to do.

What editors does CogniSpec support?

CogniSpec is a VS Code extension and can be used in VS Code or Cursor. Install it from Open VSX, open a project folder, then run CogniSpec: Create Cognitive Study from the Command Palette.

What files does it generate?

Each generated study folder includes a runnable index.html experiment, cognispec-study.json with the versioned study specification, and a README. A .gitignore helps keep participant CSV exports out of Git by default.

Does CogniSpec collect participant data or telemetry?

No. CogniSpec does not collect telemetry or transmit study specifications or participant data. Participant CSV data downloads locally in the browser.

Does validation mean a study is scientifically valid?

No. CogniSpec validates technical constraints only. A passing result does not establish construct validity, provide clinical assessment, replace piloting, or replace IRB or ethics review.

Does the generated experiment work offline?

Generated previews retrieve jsPsych from its official CDN. For an offline or locked-down deployment, use a locally pinned jsPsych copy.

Is CogniSpec open source?

Yes. CogniSpec is MIT licensed. The source, privacy documentation, and security documentation are available on GitHub.

Start with a visible specification

Build the task. Keep the reasoning attached.

Install CogniSpec, create a study from the Command Palette, and keep the configuration that produced your experiment in the same project.

Install CogniSpec View on GitHub