Research

Coming soon

Give your question room to develop.

Chat Deep Research: ask in conversation, get a sourced report with citations. Fast and limited; the full research workspace is coming.

Try it in Chat Explore Labs

Deep Research in ChatIllustrative example

Question

How will humanoid robots change factory work over the next decade?

Understand(done)
Plan(done)
Search(done)
Read(done)
Compare(done)
Synthesize(done)
Cite(done)

Plan

Q1Where are humanoid robots actually deployed today?
Q2Which tasks are technically within reach?
Q3What do labor studies expect to change, and for whom?

4 searches

humanoid robot factory pilot results
humanoid manipulation benchmark dexterity
automation exposure manufacturing tasks study
humanoid robot cost per hour estimate

Sources 6

1Field notes from a humanoid pilot on an assembly lineArticle Trade publication
2Benchmarking whole-body manipulation in unstructured settingsPaper Robotics research paper
3Task-level automation exposure in manufacturingReport Labor economics report
4Occupational task inventory, production workersDataset Public statistics dataset
5Safety requirements for collaborative robot cellsTechnical Standards documentation
6Manufacturers' outlook on humanoid adoptionArticle Industry survey write-up

Report

Timelines diverge: the survey expects broad adoption sooner than benchmark results support.

Early deployments handle material movement, not skilled assembly. 16
Dexterity in varied, cluttered settings remains the technical limit. 25
Effects arrive task by task, starting with handling and machine tending. 34

Over the next decade, expect humanoid robots to take on repetitive material handling first, while varied manipulation and safety integration limit how far they reach into skilled work.

An illustrative example written for this page — not a live run, and not real search results.Ask it in Chat

How a run works

A hard question, worked in the open.

Deep Research plans, searches, reads and compares before it writes. Step through an example run, then switch the question and watch the same shape hold.

UnderstandIndustry example

“How will humanoid robots change factory work over the next decade?”

Goal

Separate what pilots have shown from what is still projection, and say which kinds of factory work change first.

Constraints

Prefer deployment evidence over announcements
Keep labor-market effects distinct from technical readiness
Flag where sources disagree

Purpose

Keep the question at the center.

Deep Research in Chat moves a question from prompt to sourced report: run, sources, citations, final answer. A full research workspace with projects and evidence tracking is planned.

RunDeep Research plans the question and works through it in steps.
SourcesThe pages it read come back with the answer.
CitationsClaims point to the sources behind them.
Final answerA report you can read, question and follow up on in Chat.

Every claim keeps its receipt.

Open a source to see the passage that mattered, the subquestion it answers and the findings it supports. Filter by the kind of document.

[1]ArticleTrade publication

Field notes from a humanoid pilot on an assembly line

The pilot moved totes between fixed stations; tasks requiring fine assembly stayed with people.

Why it matters: Direct evidence of what a deployment does today, not what it might do.
Answers: Q1 Where are humanoid robots actually deployed today?
Supports: F1 Early deployments handle material movement, not skilled assembly.

Example source, described by kind and origin. A real run links each citation to the page it came from.

Evidence map

Follow any claim back to where it came from.

A question fans out into subquestions and sources, then narrows into findings and one answer. Point at any part to light up the path that connects it.

QuestionHow will humanoid robots change factory work over the next decade?

Subquestions

Sources

Findings

Answer

Select any subquestion, source or finding to trace the evidence path.

Questions worth the depth

For questions with more than one good answer.

Deep Research suits questions where the evidence is spread out and sometimes disagrees. Pick one to see how a run would break it down, then ask it in Chat.

Technology

Which memory designs look most promising for long-running AI agents?

What a run would look into

Q1How agents store and retrieve context over long tasks
Q2Where retrieval breaks down as memory grows
Q3What evaluations exist for long-horizon recall

Kinds of sources it would weigh

Paper, about 50 percent of the mix
Technical, about 33 percent of the mix
Article, about 17 percent of the mix

Ask this in Chat Opens Ethen Chat with this question. Turn on Deep Research there to run it.

Control

Research you can switch on, check and delete.

You decide when it runs.
Deep Research is a toggle in Chat. It runs only on the messages where you turn it on.
Sources stay with the answer.
Sources and citations are saved with the chat message, so you can go back and check what a claim rests on.
Deleting the chat removes them.
Sources live inside the conversation's messages. Delete the chat and they go with it.
Providers are named.
A search query derived from your prompt, and the pages to read, go to the search and extraction providers listed on our subprocessors page.

Identity

Research work in Ethen.

Deep Research is a Chat capability. Ethen Labs is the company research organization.

Ethen Research: Deep Research in Ethen Chat, for your own questions. Available in Chat today; a full research workspace is planned.

Ethen Labs: The company's research organization. It studies Ethen's own questions and publishes what it finds.

FAQ

Is Research the same as Ethen Labs?

Deep Research is a Chat capability for your research questions. Ethen Labs explores the company's research questions.

Which sources can Research access?

Source access depends on the supported product configuration. Broad coverage and citation guarantees are not established.

Research in Chat today.

Ask a question in Ethen Chat, or learn about Ethen Labs.

Try it in Chat Explore Labs

FAQ

Is Research the same as Ethen Labs?

Deep Research is a Chat capability for your research questions. Ethen Labs explores the company's research questions.

Which sources can Research access?

Source access depends on the supported product configuration. Broad coverage and citation guarantees are not established.

Give your question room to develop.

Plan

Sources 6

Report

A hard question, worked in the open.

UnderstandIndustry example

PlanIndustry example

SearchIndustry example

ReadIndustry example

CompareIndustry example

SynthesizeIndustry example

CiteIndustry example

Keep the question at the center.

Every claim keeps its receipt.

Field notes from a humanoid pilot on an assembly line

Follow any claim back to where it came from.

For questions with more than one good answer.

Research you can switch on, check and delete.

You decide when it runs.

Sources stay with the answer.

Deleting the chat removes them.

Providers are named.

Research work in Ethen.

FAQ

Research in Chat today.

Give your question room to develop.

Plan

Sources 6

Report

A hard question, worked in the open.

UnderstandIndustry example

PlanIndustry example

SearchIndustry example

ReadIndustry example

CompareIndustry example

SynthesizeIndustry example

CiteIndustry example

Keep the question at the center.

Every claim keeps its receipt.

Field notes from a humanoid pilot on an assembly line

Follow any claim back to where it came from.

For questions with more than one good answer.

Research you can switch on, check and delete.

You decide when it runs.

Sources stay with the answer.

Deleting the chat removes them.

Providers are named.

Research work in Ethen.

FAQ

Research in Chat today.