Frequently Asked Questions

Sefiron is an AI research agent for financial questions. You ask in plain English, and it plans the research itself: it pulls live market data, searches SEC filings, checks regulatory feeds and prediction markets, runs any calculations it needs, and returns a written answer with interactive charts and a citation for every source it used.

That could work. But then you are stuck setting up the feeds, normalizing data, plumbing through citations and replicating the work that's already gone into Sefiron. Wouldn't you rather be trading?

No. There is no query language and nothing to configure. Type a question the way you would ask a colleague, like "is Tesla expensive relative to its sector?", and Sefiron works out which data it needs. If a term in the answer is unfamiliar, ask it to explain.

No. Sefiron is a research and information tool, not a broker, adviser or registered investment adviser. It does not make recommendations, manage money or place trades, and nothing it produces is a solicitation to buy or sell any security. Every decision is yours.

• Market data — live quotes and price history, company fundamentals, options chains, analyst estimates, earnings, corporate actions and ownership
• SEC EDGAR — company filings, searched and converted to readable text
• FINRA — short interest, short volume, threshold lists and OTC/ATS activity
• Prediction markets — live odds from Kalshi and Polymarket, plus on-chain flow and trader positions
• Sports — playoff series state and league tables for the major leagues
• The open web — when a question needs something the sources above do not cover
• Your own documents — PDFs and images you upload, converted to searchable text

Quotes and price history are fetched when you ask, not cached from a previous session. Regulatory data arrives on whatever schedule its publisher uses. FINRA short interest, for example, is reported twice a month, so the most recent figure is always somewhat behind the market. Sefiron shows the as-of date for every figure.

Yes. Sefiron searches EDGAR for the filing you mean, converts the document to readable text, and summarizes or quotes the sections relevant to your question, with a citation back to the filing so you can check it. That covers 10-K and 10-Q reports, 8-K events, proxy statements and ownership forms.

Yes. Sefiron can rank and screen companies worldwide, not only US listings, and it normalizes the results to USD so a comparison across countries is meaningful. You can ask for a global ranking or narrow it to a region such as Europe or Asia.

Yes. Every answer carries citations, and the step timeline above it shows the research the agent performed: which tools it called, in what order, and what each one returned. If a figure looks wrong you can trace it back to its source.

Yes. When a question needs real computation, like a correlation, a backtest or a weighted average across a screen, Sefiron writes and runs Python then uses the result in its answer. Arithmetic in an answer is computed, not estimated.

Yes. Upload a PDF or an image, such as a fund factsheet, a pitch deck or a scanned report, and Sefiron converts it to searchable text. It can then answer questions across your documents and live market data together, citing both.

A persona tells Sefiron how you invest, so the same question gets framed for you: your risk tolerance, your time horizon, and how much technical depth you want. You can pick a built-in preset, write one yourself, or let Sefiron interview you and draft one from your answers. You can keep several and switch between them.

New accounts start with free credits and no credit card. After that there are two ways to pay. Pay-as-you-go bills you only for the research you run, and costs scale with the work: a query that calls many tools costs more than a single quote lookup. The Standard plan is $20 a month and includes a $20 monthly credit budget, along with a priority analysis queue and email support. The two work together rather than being a choice you are locked into — usage beyond a plan's monthly budget simply falls back to pay-as-you-go, so a heavy month is never a hard stop.

A team with more than twenty years building trading and research systems at major investment banks and hedge funds. We built the research tool we wanted on the desk and could not buy without an institutional contract.