Practical answers

What you need to know without opening technical details

Where do I write the question? Quick research is the fastest starting point. Open Advanced research when you want to review the full setup before launch.

Where do I find the PDF? When the document is ready, the PDF appears on Home, in My Research, on the research page, and inside the final report.

What should I do while the research is running? Usually nothing. Intervene only when the materials, constraints, or the question itself have changed.

Which page should I open every day? Open Quick research to start something new, My Research to find active or finished work, and the final report when you want to read or print.

In plain words

Why this tool matters

Most tools give one answer. This one helps you build a research path, keep evidence, and return later without losing context.

Most searches are scattered. This one groups questions, sources, notes, artifacts, and candidate results into one run.

Most AI outputs are hard to trust. This one is designed around sources, traceability, review, and human judgment.

What happens operationally

Step by step

1. Start from a question You write a question in ordinary language. AutoSearch turns it into a research path with a clear objective, source plan, and final document.

2. Choose the right entry Quick research helps you start immediately. Advanced research gives more control over workspace, depth, final document, and context before launch.

3. The model assists the research The AI Agent proposes the next useful step: a candidate idea, a mini-plan, a note, a comparison, a melody sketch, or an investigative angle depending on the lab type.

4. You review evidence, not just output Each run keeps artifacts, logs, notes, and score changes so you can judge what is useful, what is weak, and what deserves follow-up.

5. You keep, pause, or promote Nothing goes forward automatically. The human stays in charge. The lab is built to help judgment, not replace it.

Typical use cases

What you can do with it

Scientific scans across papers, trials, authors, and open questions

Public-source investigations on people, companies, events, and claims

Security research on vulnerabilities, exploit paths, and mitigations

Market and positioning studies for products, sectors, or competitors

Music ideation such as piano motifs, symbolic melodies, and MIDI sketches

Structured knowledge work: memos, summaries, literature maps, and briefings

Interesting results you can reach

Outputs worth showing to others

A structured run with objective, source plan, budget, and governance rails

Artifacts you can inspect: notes, summaries, candidate proposals, and evidence files

A report you can export and share with colleagues or students

A daily review flow that surfaces what improved, what failed, and what deserves attention

A reusable workflow that works for many types of research instead of a single niche task

What the model really does The AI Agent, in practical terms

It does not simply answer once and disappear. It helps you iterate.

It does not replace sources. It helps organize, compare, and extend them.

It does not decide for you. It proposes, scores, summarizes, and leaves the final judgment to the researcher.

Its value grows when the question is clear and the source plan is explicit.

Examples

Good first tests

Professor / research use Compare the latest papers on a topic, extract the main schools of thought, and suggest open questions worth studying.

Investigation use Reconstruct the public timeline of a company or person and check whether public claims are consistent across sources.

Security use Map a vulnerability family, summarize exploitation patterns, and propose a mitigation checklist.

Creative use Generate a piano motif, produce a symbolic note sequence, and export a MIDI sketch for review.