AstraEye Labs
Sign InEnter the Lab
AstraEye Labs · For Education

TEACH THE METHOD

Students are usually asked to defend a method they were never shown. AstraEye is built to show it. Analyses carry their methodology: the theory a method rests on, its sample-size caveats, and what it cannot support.

What a class studies

RESEARCH DESKS

Research is divided between specialist desks, each covering one discipline across US equities and digital assets. A mentor picks the desks a question needs, so a class works the way an investment committee does: several methods examined side by side, disagreeing where the evidence disagrees.

ATLAS

Technical Market Research Analyst

“Price is behavior. I map the structure behind it.”

Examines price, volume, momentum, trend, volatility, breadth, correlation, and session data across crypto and equities. ATLAS describes technical conditions and relative strength without selecting entries or exits.

Market StructureMomentumVolatility

FELIX

Fundamental & Macro Research Analyst

“Durable returns need durable reasons.”

Examines tokenomics, protocol revenue, development and governance activity, equity financials, earnings quality, estimates, valuation, and macro regimes. FELIX organizes fundamental evidence without rating an asset for purchase or sale.

FundamentalsValuation ContextMacro Regimes

NOVA

News, Catalyst & Sentiment Analyst

“Markets move when the story changes.”

Summarizes news, regulatory events, scheduled catalysts, analyst actions, earnings calls, social activity, and source quality across crypto and equities. NOVA describes narrative shifts and event context without labeling them trade setups.

News & CatalystsSentimentSource Quality

LYRA

Research Synthesis Analyst

“One finding is a note. The committee is the score.”

Compares enabled scientist findings, maps agreement and conflict, and classifies broad market regimes. LYRA summarizes research relationships without evaluating a user's decisions, thesis, or performance.

ConsensusConflict MappingMarket Regimes

SAGE

Research Summary & Explanation Interface

“Ask me about the completed research.”

Turns completed runs into final summaries, visual briefs, anticipated questions, and explanations grounded in the available scientist findings. SAGE presents research in plain language without making investment decisions for the user.

Research SummariesVisual BriefsQ&A

SCOUT

Market Discovery & Screening Analyst

“Research starts at the edge of the map.”

Builds liquid crypto and equity universes and runs standardized screens for momentum, breakouts, earnings momentum, emerging themes, new listings, IPOs, sector context, and watchlist research. SCOUT surfaces candidates for further research, not investment recommendations.

ScreeningDiscovery ResearchWatchlists

DELTA

Derivatives & Volatility Research Analyst

“Options price the distribution, not the direction.”

Examines the derivatives market: implied-volatility surfaces and term structure, put/call skew, estimated dealer gamma, options flow, perpetual funding regimes, open-interest positioning, liquidation clustering, basis and carry, and the premium implied volatility carries over realized. DELTA describes observed derivatives pricing and positioning, and does not tailor conclusions to user holdings or identify a transaction.

DerivativesImplied VolatilityPositioning & Flow

NEXUS

On-Chain & Alternative Data Research Analyst

“Capital leaves footprints before price reflects them.”

Examines on-chain flows, holder distribution, network use, stablecoin liquidity, whale activity, and staking, alongside equity short interest, borrow fees, insider activity, and institutional positioning. NEXUS reports observed flows and positioning without tailoring conclusions to user holdings.

On-Chain DataHolder & Network ActivityMarket Positioning
Going deeper than the report

ASK UNTIL IT MAKES SENSE

A finished report tells a student what a method concluded. It does not, by itself, teach them the method. SAGE is where that gap closes: a student can ask how an analysis works, why it is constructed the way it is, what its assumptions are, and where it stops being reliable.

The important part is what those answers are built from. When SAGE explains a method it is given that analysis’s own written methodology from the catalogue, so the explanation is AstraEye’s documented approach rather than the model’s general impression of an indicator with a similar name. Asked cold, a language model answers about method with textbook defaults: fluent, confident, and indistinguishable from truth to exactly the audience that most needs it to be true.

So a student can work outward from one figure they did not understand to the theory underneath it, at whatever depth they want, and what they are told stays anchored to the work in front of them.

How a class runs

FIVE STEPS

Step 01

The institution takes seats

A principal is enrolled by AstraEye, then runs the institution without us. They invite mentors; mentors invite students. Headcount is capped by seats, and a principal cannot appoint another principal, so delegation stops one level down.

Step 02

A mentor runs the research

A mentor picks the specialist desks, sets the scope and launches a run across US equities and digital assets. Principals administer the classroom rather than run research, and students never start it at all, which is why their accounts cannot reach the lab.

Step 03

Finished work is assigned

The mentor assigns a completed report to named students or to a whole class. The run belongs to the mentor; the student is granted a read of it, and that grant is checked before anything is fetched.

Step 04

The student interrogates it

Students read the findings and question them in a chat that carries the run’s full context, so the answers stay tied to the evidence in front of them. Audio narration and translation are available as options a mentor chooses when running the work.

Step 05

They sit a test, staff see the scores

A test is generated from the report itself. Scores land in a gradebook with monthly aggregates and a CSV export the institution can take on demand.

What each account can touch

THREE ROLES

CapabilityPrincipalMentorStudent
Create and run researchNoYesNo
Assign work, run classes, see the gradebookYesYesNo
Invite studentsYesYesNo
Invite a mentorYesNoNo
Read an assigned report and sit its test––Yes
Connect a brokerage or an exchangeNoNoNo
Import a portfolioNoNoNo

No role in an institution can connect a brokerage or an exchange, or import a portfolio. That is a property of the account rather than a setting. The restriction governs what an account may start, so an account that already held a connection before joining keeps it, which means a cohort issued fresh accounts reaches no personal financial data at all. It is usually the first thing a compliance office asks about.

How students work with a report

READ, QUESTION, DEFEND

Ask any paragraphEvery paragraph carries its own question mark, so a student challenges a claim exactly where it is made rather than describing which sentence they meant. SAGE replies draw on the asking account’s own plan and credits.
A console, not a chatbotSAGE docks beside the findings with the run’s full context loaded, so its answers are about this report and its data rather than the internet’s general impression of the subject.
Questions answered before they are askedEach run ships the questions its own findings invite, answered up front. It is a useful model of what a careful reader should have wanted to know.
Charts where the data supports themWhere an analysis produces chartable figures the desk renders them, so a cohort can argue about the shape of the data and not only about the sentence describing it.
A visual brief per scientistEach specialist desk produces a one-screen visual summary of what it found, which is usually where a seminar discussion starts.
The methodology, in fullWhere a run carries methodology, it gives the theory a method rests on, its sample-size caveats and what it cannot support. Some composite scores are proprietary and state the approach without the formula.
Navigate a long reportFindings are addressable by section, so an instructor can point a class at one part of a long run instead of asking thirty people to scroll to roughly the right place.
A test per deskEach scientist’s test is generated from its own findings and submitted separately, so a sitting can be part-finished and a score reflects the desks completed so far.
Listen instead of readReports can be narrated so students work through them away from a screen. Playback is an option bought with credits rather than something every account carries.
Spanish and RussianA run can be translated for cohorts who would rather read in another language. Translation is purchased with the run by the mentor, not enabled per student.
How a sitting is recorded

GRADING AND INTEGRITY

Where a student is given a report they have already sat, the repeat is recorded as practice and kept out of their average, because they have already seen the answers.

The run belongs to the mentor and the answers belong to the student. Those two identities are kept separate deliberately, so a student opening an assigned report reaches that report and nothing else of their mentor’s work.

Pilots

START WITH ONE CLASS

Institutions are set up under a written agreement, with seats for instructors and students and a price agreed per institution. If you run a student-managed fund or teach a finance, economics or business course, we would rather hear whether the shape is right than keep guessing.

Talk to us about a pilot