An independent research project
Give your financial research a memory. Connect companies, mutual funds, macro forces and crypto networks in a graph your agent can build on.
Your watchlist. Your questions. Your cadence.
Surface connections worth investigating—not another feed to scroll.
Follow a possible transmission path, then ask your agent to check the evidence.
Illustrative workflow · Not a live agent or trading signal
The project collection
Start with the market you care about.
Explore the public work, then shape a research system around your own questions.
Follow the links between policy, commodities, currencies, supply chains and the companies exposed to them.
11 nodes · 9 relationships. A source-linked policy and energy seed is ready to download. The broader private tracker is separate.
Connect Indian companies with their sectors, macro drivers, management commitments and changing fundamentals.
10 nodes · 10 relationships. Start with TCS, Infosys and HDFC Bank, plus sector, FX and lending dependencies.
Build context around filings, earnings, business drivers and the relationships behind a company’s investment thesis.
10 nodes · 10 relationships. Start with Microsoft, Amazon and NVIDIA, plus cloud and compute research questions.
Look beyond price: connect protocols, adoption, liquidity, incentives and token supply to the questions that matter.
10 nodes · 12 relationships. Start with Ethereum, ETH, Aave and USDC. The full private crypto tracker remains separate.
Connect scheme and plan choices with expenses, underlying exposures, debt risks and passive benchmarks.
12 nodes · 12 relationships. A source-linked starting map for fund research. The existing private fund tracker remains separate.
Seed v0.1.0 · Source reviews: 11–12 September 2026 · Original topology and summaries: MIT. Linked sources retain their terms. No live prices or automated monitoring. Machine-readable seed catalog ↗
A graph that works your way
We’re exploring personalized graphs built around your universe, your hypotheses and the signals you want to follow.
A sector, a portfolio, a handful of protocols or a macro question.
Keep facts, interpretations and your own scoring rules separate.
On demand, after the close or every week. You control the runner and its permissions.
For the agent at the other end
A small, inspectable starting point—not a black-box score or a promise of market-beating returns.
Record the source, observation time and what was actually reported. Missing evidence stays missing.
Store the proposed mechanism alongside its evidence. A hypothesis is not a verified causal relationship.
Use your rules to flag a change for review. Keep the supporting path visible and the final judgment yours.
The human-readable side
toroIQ began with market briefs and videos. That work is still here: what moved, what might explain it, and what deserves a closer look.
Market explainers and research videos.
Written briefs, observations and market context.
The podcast companion to the written research.
Built in the open, one piece at a time
Built by Nimit Mehra out of a personal research practice spanning Indian equities, US equities, mutual funds, crypto and macro. The aim is to make that work useful—and adaptable—to other people and their agents.
Five source-linked seed packs and a workspace initializer are open source. Start with these focused maps, or ask us to help seed a graph around your own universe. Full private tracker releases remain separate.
Yes. Download a source-linked seed for any of the five lenses. The initializer creates a local graph, configuration, source-review queue and update guide. It does not fetch live data, call a model or schedule itself. The older fictional demo is retained separately.
The goal is to notice relevant changes and overlooked connections sooner in your research process. Whether a signal is useful must be tested. No predictive accuracy, investment return or timing advantage is promised.
Yes. Start with a whole pack or an entity and its neighboring relationships, then keep your own research in your repository. The seeds include questions to investigate. Production source connectors, scoring, change detection and scheduling still need implementation.
No. There is no hosted data API, managed scheduling service or live MCP endpoint here. Public research artifacts and demo files have different scopes and should not be treated as a licensed, complete market dataset.