August 3, 2026Analytics
Portfolio risk answered the same way everywhere
A portfolio is now a first-class subject for risk, so every surface that asks what a book’s risk is gets the same answer from the same computation.
- Sleeve netting happens once, on the server, for both the portfolio view and the Aegis deep link
- The deep link no longer rounds netted weights before asking, so small positions stop disappearing into rounding on the way to the risk engine
- A strategy target the preview cannot reproduce is refused rather than shown
- Paper performance reports on an explicit model-NAV basis, with the basis labelled on the figures
August 1, 2026Platform
Strategy exports you can pin and download
Exports now pin to a specific strategy version and come back with a receipt, so the file you hand to a colleague or an allocator is one you can prove you generated. Downloads work directly from the command line and from MCP, not only in the browser.
- Every export carries its strategy version and a content hash, so two files can be compared without opening them
- Pine Script export refuses to emit a strategy it cannot render faithfully, instead of quietly falling back to an always-true entry condition
- Whether a strategy can be exported is decided by actually running the exporter, so the badge on screen cannot drift from what the file will contain
- Exporting a strategy says nothing about broker eligibility. No broker connection is live, and nothing here places an order.
August 1, 2026AI
Vega covers global indices and says when it cannot answer
Comparisons can now reach the major global indices, and when Vega runs into the edge of what it can support it tells you which limit it hit rather than answering with something adjacent.
- Global index comparisons across SPX, FTSE, DAX, Nikkei and Nifty
- Asking for a named dataset resolves to that dataset instead of something near it
- A declined question says what kind of limit it hit, so you can tell a coverage gap from an unsupported request
- Requests for figures the data cannot support are refused rather than estimated
- News evidence reads the article lane and marks absence explicitly, so a quiet week reads as quiet instead of as nothing found
July 29, 2026AI
Vega reads your saved portfolios
Ask Vega about a portfolio you have already saved and it reads the real book rather than inventing a stand-in. Named portfolios go straight to the risk view, and Vega can show the same factor decomposition the Aegis cockpit shows.
- Saved portfolios from both Vega and Aegis are readable by name in a conversation
- Factor risk and the top ten risk contributors come through in chat, from the same computation the cockpit uses
- Charts are materialised on the server at full resolution from the query Vega already ran, so a long-window chart is not thinned down to fit a message
July 27, 2026Analytics
Aegis reports you can hand to a client
The risk cockpit now exports work you can put in front of someone else. The old raw CSV dump is replaced by a structured export, and there is a detailed vector PDF alongside it.
- Structured CSV built to be read and interpreted, rather than a flat dump of every number
- Vector PDF report with TradePolaris and Aegis branding
- Methodology, confidence, concentration, factor, correlation, drawdown, liquidity, options and caveat sections, so a reader can see how each figure was produced
- Both download straight from the cockpit
July 27, 2026Platform
Deployment reports and signals that expire
A deployed strategy can now account for itself. Generate a report on any deployment on demand, and every signal carries an explicit expiry so an old one is never delivered as though it were current.
- On-demand deployment report, generated when you ask for it rather than on a schedule
- Every signal is stamped with an expiry when it is emitted, using market-aware arithmetic that accounts for sessions, weekends and holidays
- A signal past its expiry is dropped at delivery instead of arriving late and looking fresh. The arithmetic errs toward expiring early, never late.
- A guided path to a first deployment, and the signal feed can be copied or exported as CSV
- Deployments remain simulated. There is no broker connection and nothing places an order.
July 19, 2026Analytics
Portfolios hold options and futures, not just shares
A book can now carry option legs and futures positions alongside its equity, and those positions flow through to risk rather than sitting on the side as a note.
- Option legs on sleeves, folded into live risk instead of tracked separately
- Futures positions with the correct CME contract multipliers
- Describe holdings in plain English on the import screen, or re-import into a portfolio you already have
- Option legs carry end to end, into the Aegis cockpit and into what Vega reads in chat
- Factor-model covariance for large books, so a several-hundred-name portfolio still gets a real risk model
July 19, 2026Platform
A strategies library, and a builder that keeps your work
Strategies live in a library rather than scattered across runs, and the builder no longer loses a draft you were part way through.
- A library of your strategies, with a stepper for starting a new one
- Draft recovery, so a half-finished strategy survives a closed tab
- Progressive disclosure in the builder, so the full surface does not arrive all at once
- A full walk-forward run asks you to confirm the credits before it starts
July 18, 2026AI
Saved apps live on the dashboard and refresh overnight
Apps you save now have somewhere to live. They appear on your dashboard, collect on a single board, and re-run themselves after the close so what you see in the morning is current.
- Saved apps on the dashboard, and an unfinished draft is restored in the builder at no cost
- A board showing every saved app's snapshot on one grid
- Nightly auto-refresh after the close, run by the worker rather than on your next visit
July 18, 2026Data
Live quotes, and charts that remember where you were
Tier-1 quotes now stream live through a conflated fan-out, with a watchdog on feed freshness. The professional TradingView charts are armed, and the asset page stops throwing away your view.
- Live streaming quotes with a freshness watchdog on the feed
- Professional TradingView charts with chart state remembered per account
- The asset page no longer jumps back to the top when the chart loads, and the first bars paint sooner
July 15, 2026Data
More of the market in the data lake
Several lanes were added to the lake this week, across US short interest, factor research and Indian derivatives.
- FINRA short interest and short volume, IPOs, and a deeper equities-minute lane
- Ken French factor lane with a weekly sweep
- NSE futures and options universe covering 159 names, plus 18 global and NSE index lanes
July 13, 2026AI
Vega Apps — living dashboards from a sentence
Describe a chart or monitor in plain English and Vega builds a small app you can save, refresh, edit and share. Start from the template gallery or from scratch, refine it in conversation, and turn any single-symbol app into a live alert.
- Template gallery — ready-made apps you can clone and rebind to your own symbols (no credits to rebind)
- Refine in conversation: “make it weekly” edits the same app instead of starting over
- “Watch this” turns an app’s condition into a conditional alert delivered to your Updates inbox
- Public share links — anyone can view, and fork a shared app into their own account
July 13, 2026Platform
New plans: Pro, Aegis, and Funds & RIAs
Pricing now matches how people use the platform. Pro ($49/mo) is for serious, recurring research. Aegis ($249/mo) adds the institutional risk suite with a 5,000-credit monthly grant. Funds & RIAs run Aegis institutionally from $1,000/mo. Charts and market data stay free for everyone, and every free account gets 200 credits monthly.
- Pro — 1,000 credits/mo, 10 paper-trading slots, 10 saved Vega apps
- Aegis — the full risk cockpit, plus everything in Pro and 20 saved Vega apps
- Funds & RIAs — multi-book portfolios, onboarding & book migration, LP-ready reporting
July 12, 2026Analytics
Aegis — the institutional risk cockpit
Aladdin-grade risk for your own book, recomputed live as you edit it. A tabbed cockpit puts ex-ante VaR, stress testing, factor risk and the optimizer one click apart, with deep links to every view and plain-English risk narratives grounded in the numbers on screen.
- Ex-ante VaR & CVaR, factor decomposition, exposure & concentration — live as the book changes
- Scenario & stress testing: historical crisis replays, macro & factor shocks, Monte-Carlo
- Institutional optimizer — sector/turnover constraints, Black-Litterman views, and a never-worse minimum-variance mandate
- Multi-strategy sleeve portfolios with aggregate risk
July 2026Data
Global search: the full 13F universe, politicians & activists
⌘K search now covers every institution in the recent 13F filing directory — thousands of managers, with accession-level positions loaded on demand — plus politicians’ trades and 13D/G activist stake filers, all searchable by name.
- Search any indexed 13F filer by name and jump straight to its latest filing accession
- Politicians and 13D/G stake filers resolve by entity name from the same search bar
June 2026AI
Vega — your AI research analyst
Vega is a chat analyst that does the work, not just the talking. It queries the data lake, searches the web for context, drafts a strategy, and runs a preview backtest — showing each step live with the SQL and sources cited.
- Author charts and tables on the fly from a plain-English request
- Citation cards: every answer pairs a source, a chart, and the query behind it
- Proposes the big actions (full validation, paper deployment); you confirm — it never auto-executes
- @-mention any stock or dataset to pull it into the conversation; a built-in cost meter keeps spend visible
June 2026Analytics
Portfolio & holdings analytics vs the S&P 500
Backtest reports now chart the equity curve against an S&P 500 benchmark and break performance down by holdings, so you can see where return and risk actually came from relative to the market.
- Benchmark overlay on the equity curve with alpha, beta, R², tracking error and information ratio
- Up/down capture and a full Fama-French 5-factor + Momentum decomposition
June 2026Data
Live news feed + latest headlines
A live news feed with per-ticker sentiment now runs across the terminal. Ask the data for the latest headlines on a name and get a current article stream alongside the numbers.
- Per-ticker sentiment with an AI-scored overlay
- Inline on asset pages and queryable in plain English
June 2026Workflow
Conditional alerts + an Updates inbox
Set conditional alerts on the data you care about and receive them in an in-app Updates inbox — so a watched condition or a strategy event reaches you without babysitting a screen.
June 2026Engine
Honest, survivorship-aware backtests
Our backtest engine uses historical membership when that lane is available, retains later-delisted names where supplied, and warns when a run falls back to current constituents. It also models realistic trading costs. We would rather show a lower, qualified number than a flattering one. Backtested results are hypothetical and not indicative of future results.
- Delisted-symbol universe option on the full tier, with universe-mode warnings
- Next-bar-open fills, gap-aware stops, bid/ask spread, square-root market impact and short-borrow costs
- Walk-forward, out-of-sample validation — not a single in-sample curve fit
June 2026Data
Index & sector screens
Screen by index membership (S&P 500, Nasdaq-100, Dow) and by sector with curated, point-in-time data — no live web round-trip needed for the common universes.
- Real index levels and ~3 years of price history across 11 indices
- Sector heatmap, breadth and most-active names on the live markets dashboard
June 2026Analytics
Institutional tearsheet & on-demand analytics
Every backtest produces a professional report, and you can run the same metric suite on any ticker, basket, or benchmark on demand — no backtest required.
- Performance: total return, CAGR, Sharpe, Sortino, Calmar, monthly-returns heatmap
- Risk: volatility, max drawdown + longest underwater stretch, Ulcer index, VaR & CVaR (95%/99%)
- Overfitting credibility layer: Probabilistic & Deflated Sharpe, Minimum Track Record Length, Probability of Backtest Overfitting
June 2026AI
Deep Research mode for Ask the Data
For questions that name a universe we can't list directly ("the Fortune 500 by net margin"), Vega does a live web search to resolve membership only, then pulls the actual figures from licensed data — so the answer is vendor-grade, not scraped guesswork.
June 2026Platform
Usage-based credits & shareable tearsheets
Pay only for the compute you run. Mint a public link to any backtest report to share what you built, with a clean social preview card.
- 100 credits = US$1; 200 free credits every month, no card
- Ask the Data and preview backtests are free; a full walk-forward backtest is priced on its actual compute — from ~96 credits ($0.96), scaling with history, factors & positions
- Shareable public tearsheets + a referral waitlist with line-jumping
TradePolaris is research software, not investment advice. Backtested and hypothetical results do not predict future performance. Trading is paper (simulated) only. See our Disclaimer.