Getting started2 min readUpdated 2026-11-09

How to make one project, step by step

A project takes a handwritten concept to a first manuscript draft in six steps: Input, Novelty check, Code generation, Code execution, Manuscript and Delivery. You create an account, connect your own Claude subscription, buy a credit, create a project, and then follow the real screens below.

This guide takes you from nothing to your first running project. Allow about 15 minutes for setup, plus the time each stage needs.

Before you start

  • A Windows 10 or 11 computer
  • A Claude subscription, Pro or higher, on your own account
  • Your handwritten concept as a photo, scan or document, and your base papers
  • Your dataset, if you have one (simulation work needs none)

Walkthrough

  1. 1

    Open the app and look at the sample projects

    The home screen lists sample projects and your own. A Getting started checklist on the right confirms your account, verified email, connected Claude and first project.

    TipOpen a sample first. It shows every step after it ran.

    EasySeer Sample projects screen listing three finished example projects, each with an Open button.
  2. 2

    Create a new project in five short steps

    Name the project, choose its folder on your computer, choose the project type (Article, original research), choose the style (for example Engineering) and review.

    TipLiterature review is shown as coming soon.

    EasySeer New project screen, step 3 of 5: choose the project type, with Article (Original) selected.
  3. 3

    Add your concept, proposal, papers and data

    Upload the handwritten concept as photos, scans or PDF. Add a proposal or abstract, base papers and datasets if you have them. The more you add, the better the novelty score.

    TipOnly the concept is required. Everything is copied into your project folder.

    EasySeer Input screen for uploading the concept pages, an optional proposal and a dataset.
  4. 4

    Start the novelty check and let it run

    A progress list shows what EasySeer is doing: finding the files, reading the content, scoring each criterion, drafting the spec. You can leave the page; the run keeps going.

    TipUse the Chat button to ask what a step is doing.

    EasySeer Novelty check in progress, with a live list of steps and a Stop run button.
  5. 5

    Read the score and approve the project spec

    Each input is scored with a reason, then combined into a final score against your own threshold. Open context.docx, answer the open questions, then approve it or request changes.

    TipA score below your threshold is a prompt to strengthen the idea, not a verdict.

    EasySeer novelty result: a final score of 71 out of 100 against a threshold of 75, with a score for each input and buttons to approve or request changes.
  6. 6

    Review the generated code

    The code is written as separate files with a main program. Open any file, read it and ask the chat what it does.

    TipRead before you run. You are responsible for the method.

    EasySeer code screen: 25 generated files, 23 Python modules, with main.py open in the viewer.
  7. 7

    Run it and collect the results

    The code runs on your computer. Graphs, tables and sample outputs are saved with their file paths. Click a graph to enlarge it.

    TipCheck that every result makes sense for your data.

    EasySeer results screen after a run: 14 graphs, 13 result tables and 6 sample outputs, with example graphs.
  8. 8

    Choose the IEEE mode and tone

    Pick Journal (IEEE Transactions) or Conference, and Formal or Plain and direct tone. Optionally add the IEEE LaTeX source, an AI-use statement and a cover letter draft. Every claim is checked against your results.

    TipSwitch the AI-use statement on if your journal or university requires disclosure.

    EasySeer manuscript setup screen: choose layout, citation style and tone, and include an AI-use statement and cover letter.
  9. 9

    Read, verify and rewrite the draft

    The draft is assembled from your results, with a section outline, equations, algorithm steps and figures. Check every number and reference, then make it yours.

    TipNext come formatting and delivery.

    EasySeer manuscript screen with a section outline on the left and the method text, equations and algorithm steps on the right.

Tips

Start with a small, clear concept for your first project

Read each stage's output before approving

Keep your project folder somewhere you can back up

Common mistakes

  • Approving stages without reading them
  • Uploading a blurry photo of the concept
  • Forgetting to check that the dataset path is correct

Questions

How long does one project take?

Setup takes about 15 minutes. Typical run times on a laptop are about 2 minutes for Input, 8 for the Novelty check, 20 for Code generation, 25 for Code execution, 15 for the Manuscript and 20 for Delivery. Large datasets and higher Claude effort take longer.

Do I need my own data?

Not to try it. Each input has a Use sample files button. For simulation work you can leave Datasets empty and the code generates its own data.

Where do my files go?

Into a project folder on your computer, by default under Documents\EasySeer.

On the blog: The research paper workflow: from idea to submission in 7 stages (8 min read)

Still stuck? Ask in the tool's chat panel, or email courses@skilldux.com.

Last updated 2026-11-09. EasySeer is AI. It creates only an initial draft from the novel concept you provide, and it is designed to follow academic ethical guidelines.