Build Financial Analysis Workflows in Codex on Daloopa’s Verified Data
Codex is OpenAI’s cloud-based AI environment built for complex, multi-step work. Unlike a simple chat interface, Codex runs tasks in the background and is designed to produce structured outputs like models, documents, and analysis reports. What it needs to produce institutional-grade financial analysis is a reliable data source. Without a verifiable data source, Codex relies on the public web, where company financials are inconsistently formatted, hard to verify, and slow to parse with a high risk of extraction errors.
How Daloopa’s MCP Works in Codex
The Model Context Protocol (MCP) is an open standard that allows AI systems to connect to external data sources and tools. When Daloopa’s MCP is connected to Codex, it gives the platform direct access to Daloopa’s financial data layer: standardized KPIs, earnings filings, press releases, stock prices, 10-Ks, 10-Qs, 8-Ks, and historical financials across public companies. You tell Codex what you want to build, Daloopa provides the data, and the rest of the analysis is done for you.
MCPs make LLMs more accurate and efficient by replacing open-ended web searches with direct, structured data retrieval. Instead of Codex searching and reconciling from raw sources, Daloopa’s MCP returns exactly what is needed in a few targeted tool calls, with every figure sourced from company filings and documents and hyperlinked to the exact location in the source document.
What You Can Build with Daloopa Data in Codex
There are two ways to leverage Daloopa’s data in Codex. First, you can query Daloopa in natural language, asking for a company’s last four quarters of gross margin, pulling guidance from a specific earnings call, or searching across filings for a metric, and Codex retrieves it through the MCP. Second, you can also use any of Daloopa’s 20+ prebuilt Skills, which are specialized instructions for specific financial analysis tasks.
- Tearsheet: Generates a one-page company overview covering key financials, valuation metrics, and business context. A fast-starting point for a new name, a client call, or a coverage initiation.
- Earnings Review: Full earnings analysis with guidance tracking, comparing reported results to prior guidance and surfacing beat and miss patterns across key metrics. During earnings season it gives you a repeatable format, so you are not rebuilding the analysis structure for every name you cover.
- DCF: Discounted cash flow valuation with sensitivity analysis, grounded in Daloopa’s fundamental data as the historical base. You start with reported actuals and adjust assumptions from there, with the sensitivity table included across discount rate and terminal growth inputs.
- Comps: Trading comparables with peer multiples and implied valuation ranges across a user-defined peer group. Pulls market data alongside standardized fundamentals from Daloopa so the comparison is consistent across tickers without manually sourcing data.
- Inflection: Auto-detects the biggest metric accelerations and decelerations across a company’s full KPI set. Useful for quickly identifying where the story is changing before a meeting, a model update, or an earnings call.
- Research Note: Generates a professional Word document research note grounded in Daloopa data, structured for coverage initiation, client distribution, or internal use.
Other skills in the library include capital allocation analysis, unit economics, precedent transactions, industry comparisons, guidance tracking, IB pitch deck creation, and more.
Why Analysts Use It
The core problem is that general AI tools scrape from the public web, where financial data is inconsistent and hard to verify. Every number Codex retrieves through Daloopa MCP is sourced directly from company filings and documents, with each fundamental data point hyperlinked to the exact location in the source document.
The second thing it changes is where analysts spend their time. A meaningful portion of research work goes into tasks that precede the actual judgment call: pulling data, reformatting it, building a model shell before you can stress-test assumptions, assembling a comp set before you can draw a conclusion. When Codex handles retrieval through Daloopa’s dataset, that preparation compresses. The analyst starts at the point where their expertise is relevant: interpreting the numbers, not assembling them.
In practice, the combination covers most of the analysis cycle in a single Codex session. Earnings prep the night before, earnings review after the print, a model refresh from the latest reported figures, a comp set for an initiation, a tearsheet before a client call. For teams evaluating AI for research workflows, this combination addresses the core limitation of general AI tools for institutional use: powerful reasoning without reliable, sourced financial data.
Get Started
Ready to connect verified financial data to your LLM? Connect Daloopa MCP to Codex, with setup documentation here.
