Frequently asked questions

Arb Insights is an alternative-data engine for institutional investors. It cleans, normalizes and aligns dozens of dataset families, then delivers them three ways: a web platform, a quant feed, and an MCP connector that runs analytics inside AI tools like Claude and ChatGPT.

Three ways, all on the same normalized engine: a no-code web platform for point-and-click research, a point-in-time quant feed for your own models, and an MCP connector that brings live analytics into your AI client.

Institutional investors, including hedge funds, single-manager funds and allocators, plus data providers who want to reach those investors without exposing raw files.

Yes. Data is stored as it was released, so you can track restatements and run backtests without hindsight bias.

It is an MCP connector that puts the full Arb Insights engine inside your AI client. You run earnings previews, QTD reads, cross-dataset composites and more in plain language, and every figure is computed in a deterministic tool before it reaches the model.

MCP stands for the Model Context Protocol — the open standard for connecting AI clients to external tools and data. The connector speaks MCP, so you ask for analysis in plain language in any MCP-compatible client (Claude, ChatGPT, VS Code), and it resolves the datasets, tickers and timeframes you ask about, computes every figure deterministically, and returns results your AI can chart, combine and build on.

Any MCP-compatible client, including Claude, ChatGPT and VS Code. If it speaks MCP, it can query Arb Insights.

Language models estimate numbers; Arb Insights computes them. Calculations run through deterministic tools, so the figures your AI works with are exact, reproducible and auditable. From there your AI can do whatever the task needs — combine datasets, build comparisons, screen, chart and reason across results — on numbers it never had to guess.

Almost any analysis you can describe. Pull KPIs, fundamentals and model estimates versus consensus; track QTD and rolling growth; screen and build custom universes; layer in short interest, prices and macro series; assemble earnings previews or whole thematic baskets — then combine any of it into a single view. More than 20 tools cover the full engine, and because your AI composes them, you are not limited to a fixed menu of outputs.

Add the Arb Insights connector to your MCP client once, with no data plumbing on your side, then start asking questions. Book a demo and we will connect it on your own watchlist.

Alternative data is information outside traditional financial statements — card spend, web traffic, app usage, foot traffic, search trends and more — that investors use to gauge company performance ahead of official reports. Arb Insights normalizes and aligns these datasets so they can be combined.

Dozens of families across consumer, corporate and macro, including credit and debit card spend, e-receipts, e-commerce, foot traffic, mobile app, clickstream, search trends, company fundamentals, employment, IT spend, supply chain and macroeconomic series.

Every dataset is cleaned, aligned and stored point-in-time, so series line up out of the box and backtests stay free of hindsight bias.

Yes. You can load proprietary datasets into the platform and work with them alongside public, private and macro sources, all aligned under one ontology and ready to combine.

A no-code workspace for exploring the full Arb Insights engine. You build tables, charts and dashboards around the names and metrics you track, with no engineering required.

No. The platform is drag-and-drop, with pre-built pre-earnings and KPI dashboards you can tailor. The engine handles normalization underneath.

Yes. Send any view to CSV or Excel, and get notified in-app or by email when a watched series crosses a threshold you set.

A normalized, point-in-time data feed for your own models and backtests. Every dataset is pre-aligned, so it slots into your pipeline without bespoke cleaning.

Clean daily, weekly, monthly and quarterly series, consistent identifiers across datasets, and Arb model KPI forecasts alongside reported fundamentals and consensus.