Economic Intelligence Retrieval Toolkit / LLM Connector

A robust TypeScript Software Development Kit engineered for effortless interaction with the Economic Data Service API, featuring intrinsic interoperability with sophisticated language models through the Vercel AI SDK.

Core Capabilities

• Full TypeScript schema definitions covering every endpoint of the Economic Data Service API
• Native integration layer supporting LLM function invocation via the Vercel AI SDK
• Comprehensive fault tolerance mechanisms and strict input validation
• Intuitive, descriptive API surface with precise type enforcement and integrated documentation

Setup Instructions

# Via npm
npm install fiscal-data-retriever-toolkit

# Via Yarn
yarn add fiscal-data-retriever-toolkit

# Via Bun
bun add fiscal-data-retriever-toolkit

Primary Application Example

import { initializeFinancialTooling } from "fiscal-data-retriever-toolkit";

// Instantiate the data querying mechanisms using your access credential
const dataAccessTools = initializeFinancialTooling("your-secret-api-key");

// Invoke the tools directly
async function fetchAppleQuarterlyEarnings() {
  const report = await dataAccessTools.retrieveIncomeStatements.invoke({
    identifier: "AAPL",
    timeframe: "quarterly",
    count: 5
  });

  console.log(report);
}

LLM Interfacing (Vercel AI SDK)

The toolkit is architected for fluid operation alongside Generative AI agents utilizing the Vercel AI SDK's capacity for tool invocation.

Deployment within Next.js App Router

// app/api/conversation/route.ts
import { StreamingTextResponse, ChatCompletionRequestConfig } from "ai";
import { initializeFinancialTooling } from "fiscal-data-retriever-toolkit";
import OpenAI from "openai";

const client = new OpenAI({
  apiKey: process.env.OPENAI_API_KEY
});

// Setup data interaction layer
const dataAccessTools = initializeFinancialTooling(
  process.env.FINANCIAL_DATA_API_SECRET_KEY! 
);

// Define the functions available to the language model agent
const availableFunctionSpecs: ChatCompletionRequestConfig['tools'] = [
  dataAccessTools.retrieveIncomeStatements,
  dataAccessTools.retrieveBalanceSheets,
  dataAccessTools.fetchStockQuote,
  dataAccessTools.queryCompanyProfile
];

export async function POST(req: Request) {
  const { messages } = await req.json();

  const response = await client.chat.completions.create({
    model: "gpt-4-turbo-preview",
    messages,
    tools: availableFunctionSpecs,
    stream: true
  });

  return new StreamingTextResponse(response.body);
}

Integration with Anthropic's Claude

// app/api/conversation/route.ts
import { StreamingTextResponse } from "ai";
import { initializeFinancialTooling } from "fiscal-data-retriever-toolkit";
import Anthropic from "@anthropic-ai/sdk";

const anthropicClient = new Anthropic({
  apiKey: process.env.ANTHROPIC_API_KEY
});

// Setup data interaction layer
const dataAccessTools = initializeFinancialTooling(
  process.env.FINANCIAL_DATA_API_SECRET_KEY! 
);

export async function POST(req: Request) {
  const { messages } = await req.json();

  const response = await anthropicClient.messages.create({
    model: "claude-3-opus-20240229",
    messages,
    tools: [
      dataAccessTools.retrieveIncomeStatements,
      dataAccessTools.retrieveBalanceSheets,
      dataAccessTools.fetchStockQuote,
      dataAccessTools.queryCompanyProfile
    ],
    stream: true
  });

  return new StreamingTextResponse(response);
}

Tool Index Reference

The SDK exposes the following callable instruments:

Corporate Financial Statements

Function Name	Purpose
retrieveIncomeStatements	Fetch profit and loss statements by security code
retrieveBalanceSheets	Obtain balance sheet data by security code
retrieveCashFlows	Get statements detailing cash movements by code
fetchComprehensiveStatements	Retrieve all primary financial reports for a symbol
getRevenueBreakdowns	Acquire granular, segmented revenue figures
searchStatements	Query financial reports using definable criteria
locateMetric	Search for specific accounting line items across records

Market Intelligence

Function Name	Purpose
fetchStockQuote	Get the current market valuation for a ticker
queryHistoricalPrices	Retrieve time-series price information by code

Entity Metadata

Function Name	Purpose
queryCompanyProfile	Retrieve organizational background and core details for a code
fetchRegulatorySubmissions	Access official regulatory filings (e.g., SEC) for an entity
getInsiderActivity	Pull records of internal stock transactions (buys/sells)

Derivatives Data

Function Name	Purpose
getDerivativesBook	Retrieve current trading data for options contracts on an asset

Key Performance Indicators (KPIs)

Function Name	Purpose
getKpiSummary	Fetch an instantaneous summary of critical financial ratios and metrics for a code

Illustrative LLM Context Prompts

When configuring agents to utilize these instruments, context regarding their capabilities is beneficial. Examples follow:

For Multi-Instrument Comparative Analysis

You are equipped with data querying mechanisms for market figures and corporate specifics.
Undertake an evaluation of 'Apple' (AAPL) using the available utilities:
1. Retrieve the corporate profile details
2. Ascertain the latest trading price
3. Acquire the last three fiscal years of annual income statements
4. Synthesize observations regarding growth trajectories in sales, net income yields, and overall fiscal stability

For Sector Performance Benchmarking

Contrast the fiscal achievements of 'Tesla' (TSLA) versus 'Ford' (F) across the previous two reporting periods.
Employ the designated tools to fetch:
- Annual income reports
- The KPI summary
- Current market capitalization figures

Subsequently, articulate a comparative assessment focusing on growth velocity, profitability margins, and solvency health.

For Technical Investment Due Diligence

I am assessing a potential position in 'Microsoft' (MSFT). Can you execute the following steps:
1. Retrieve daily trading data covering the last 90 trading days
2. Analyze the resulting price momentum
3. Review the most recent quarterly P&L statement
4. Examine the KPI summary data
5. Formulate an investment recommendation based on technical movement and underlying financial strength

Advanced Configuration Options

Custom Service Root

To interface with an alternative backend endpoint:

const dataAccessTools = initializeFinancialTooling(
  "your-secret-api-key",
  "https://alternative-data-source.net/v1"
);

Response Error Management

All functions incorporate pre-configured error handling routines:

const data = await dataAccessTools.retrieveIncomeStatements.invoke({
  identifier: "NULLED_TICKER",
  timeframe: "annual"
});

if ("error" in data) {
  console.error(`Failure Code: ${data.error} - Detail: ${data.message}`);
} else {
  // Proceed with successful data processing
  console.log(data.statements_array);
}

Development Workflow

Building the Library

# Install project dependencies
bun install

# Compile the library source code
bun run build

# Generate client bindings from OpenAPI specification
bun run generate-client

Licensing

Proprietary