We built the tool we kept wishing we had.

Why TellHawk exists

Serious research usually means twenty browser tabs. SEC filings in one place, insider trades in another, economic releases scattered across a dozen government sites, and somewhere in the middle, a backtest someone posted online with no trading costs and no way to check it.

We got tired of it. Not of the work, which is the interesting part, but of the stitching together, and of never quite knowing how much to trust a result. So we built one place that gathers the official records every night and lays them out with the date and source on every figure. Then we built the Strategy Lab for the same reason: before you trust an idea with real money, you should know whether it still holds up once trading costs, hindsight and luck are counted.

That's the whole idea. Show the evidence, show where it came from, and be honest about where it falls short.

How we check our data

Official data is loaded overnight, after the US close. Then we check a random sample of what the terminal shows against the original publishers: over 680 automated checks against source documents every night. The pass rate is published in the terminal every day, failures included. Every figure carries its date and its source, so you can open the original and see for yourself.

Strategies get the same suspicion. An idea in the Strategy Lab has to get through six stages, which hold nine individual tests:

1. No hindsight. Every decision sees only what was known that day, with filings counted from their filing date.
2. Costs. Every trade is priced net of fees and slippage, then re-priced at three times the cost to see whether the edge was just cheap fills.
3. Unseen years. The latest years are held back, and the rule has to beat the market there, not only where it was tuned.
4. Luck. A statistical test, and a check that a few lucky periods aren't carrying the whole result.
5. Tries. The result is discounted for every version you tried, so the best of a big pile doesn't pass by accident.
6. Forward. Promising picks are frozen on a date and compared with the market as real days arrive.

A rule reaches "passed the tests" only if it clears all of them, and "holding up" only after its picks beat the market on new data. Backtests replay history after estimated costs. They are not a forecast. The Strategy Lab page has the detail, and the status page shows how the system is doing.

What it is and what it isn't

TellHawk is research, not investment advice. The research terminal places no orders and never connects to your brokerage account. TellHawk Trader is a separate, lighter product; it sends orders only to your own brokerage, and only after a risk check. We never hold your money.

Who it's for

• Individual investors who want to check a stock properly before buying it, and find out whether a strategy they read about online actually works.
• Equity and research analysts who want filings, ownership, insider activity and news reactions in one place, with sources they can cite and exports they can use.
• Small funds and advisers who want a testing discipline they can show a client: what was tested, how, and what failed.

We're opening access in small groups and reading every application.

Free is open to everyone and takes no application. Pro adds the deeper versions of the same screens.

Security and privacy

A few promises we're glad to make:

• We never sell your data. We don't rent it, and we don't share it for advertising either. The privacy policy has the details.
• We never hold your money. Payments go through a payment processor, and orders go from Trader to your own brokerage after a risk check. The research terminal places no orders at all.
• Encrypted connections. Everything between your browser and TellHawk travels over HTTPS.
• No passwords to lose. Sign-in is a one-time code sent to your email, so there is no TellHawk password to leak or reuse.
• Saved keys are stored encrypted. If you save an AI provider key for the optional assistant, we store it encrypted and use it only to answer your questions.
• Backups. We keep backups of our data, including off-site copies, so one failure doesn't take your work with it.

Where the data comes from

Mostly official public records. Every screen names its sources and the date the data describes.

• SEC filings: Form 4 insider trades, 13F fund holdings, 13D and 13G stakes, 8-K events, company financials
• Short-interest reports
• CFTC positioning
• Federal Reserve and FRED
• Bureau of Labor Statistics
• Bureau of Economic Analysis
• US Treasury
• European Central Bank
• FDA and clinical-trial registries
• Other government and international statistical publishers

Alongside the official records we use third-party market prices and news headlines. Those can be delayed or contain errors, so check anything important against the original source.

What we're building next

Themes, not promises of dates. We'd rather ship something right than ship it on a deadline.

• More official data sources

  Further public filings and government releases, held to the same date-and-source rule.

• Deeper strategy testing

  More ways to build a rule and stress-test it before it meets real money.

• More broker connections

  Link more brokerages for portfolio import and, where we can verify them, order routing behind the risk check.