---
title: "Should I embed fund factsheets in a vector store or call an API at query time?"
description: "Why fund figures should be fetched by ISIN at answer time rather than embedded from PDFs, and how to combine an API tool with retrieval over your own documents."
canonical: https://fundfactsapi.com/answers/fund-data-for-rag-and-llm-apps
published: 2026-09-12
updated: 2026-10-13
publisher: FundFacts API
---

# Should I embed fund factsheets in a vector store or call an API at query time?

**Short answer:** Call the API at query time. Fund figures are tabular, change monthly and must carry an as-of date; chunked PDF embeddings lose the table structure, mix old and new versions and return the wrong share class. Keep retrieval for your own unstructured documents (research notes, client files) and add a tool that fetches the ISIN's current JSON from GET https://fundfactsapi.com/api/v1/funds/{isin}, so numbers in the answer are exact and dated.

## Why embeddings are the wrong tool for factsheet figures

A factsheet is mostly tables: holdings, sector weights, calendar returns, cost lines. Chunking it for a vector store breaks rows from headers, so "5.48" ends up near "NVIDIA" by luck. Fund houses publish a new version every month, so the store fills with near-duplicate chunks from different dates and the retriever cannot tell which is current. And the same fund has several share classes with separate documents whose text differs by one letter in the name; similarity search returns the wrong one routinely.

The result is confident answers with stale or mismatched numbers, which in a financial product is worse than no answer.

## What works: structured data as a tool

Give the model a function that takes an ISIN and returns the current JSON. The payload is already structured (`headlineMetrics.ter`, `topHoldings[]`, `sector[]`, `calendarReturns`, `dataAsOf`), so the model reads exact values with a date and never has to parse a PDF.

```python
# pip install git+https://github.com/fundfactsapi/fundfactsapi-python
from fundfacts import FundFacts
ff = FundFacts()  # reads FUNDFACTS_API_KEY

def fund_context(isin: str) -> str:
    """Compact, dated context block for the prompt."""
    f = ff.get_fund(isin)
    d = f["data"]
    lines = [f"{f['name']} ({isin}), data as of {d['dataAsOf']}",
             f"Category {d['profile']['category']}; risk {d['riskRating']}/7; TER {d['headlineMetrics']['ter']}; size {d['keyFacts']['aum']}",
             "Top holdings: " + ", ".join(f"{h['name']} {h['weight']}%" for h in d["topHoldings"][:10]),
             "Sectors: " + ", ".join(f"{s['label']} {s['weight']}%" for s in d["sector"][:6]),
             "Calendar returns: " + ", ".join(f"{y} {r}%" for y, r in zip(d["calendarReturns"]["years"], d["calendarReturns"]["fund"]))]
    return "\n".join(lines)
```

Inject that block into the prompt (or expose it as a tool and let the model call it), and keep the vector store for the material that genuinely is unstructured: your research, meeting notes, client correspondence.

## Where documents still matter

For a fund the API does not cover, `POST /extract` (Pro and above) turns any KID or factsheet PDF into the same `data` shape, so your application handles both cases with one code path instead of falling back to raw text.

## Guardrails worth adding

- Instruct the model to quote `dataAsOf` with every figure and to say "not disclosed" for empty fields.
- Validate ISINs before calling (12 characters, check digit) so the model cannot invent one that happens to parse.
- Cache by ISIN until `expiresAt`; the figures do not change more than once a day.

The [fund data for AI agents and LLMs guide](https://fundfactsapi.com/blog/fund-data-for-ai-agents-and-llms) covers evaluation and prompt patterns in more depth.

## FAQ

### Do I need a credit card to try it?

No. The Free plan is 5 lookups per month with no card; sign up with an e-mail address or Google and the key is shown in the dashboard.

### Which ISINs work?

Any share class of a UCITS fund, ETF, money-market fund or open-end fund whose management company publishes a product page, factsheet and KID. Each share class has its own ISIN and is looked up on its own. Stocks, bonds and indices are not funds and return 404 fund_not_found.

## About FundFacts API

FundFacts turns any fund or ETF ISIN into one dated factsheet (fees and KID costs, the 1–7 risk indicator, holdings, sector and country exposure, returns) read from the fund house's own documents: as a client-ready PDF under your firm's name for advisers and wealth managers, as JSON through its API for developers, or inside ChatGPT and Claude through its MCP server.

FundFacts works inside ChatGPT and Claude as a connector (a remote MCP server): ask about any fund or ETF by ISIN, ticker or name and the assistant answers from FundFacts' dated figures, fees, risk, holdings, exposure and returns, with fund cards, side-by-side comparisons, holdings overlap and portfolio look-through.

- Free API key, no card: https://fundfactsapi.com/signup
- Documentation: https://fundfactsapi.com/docs (OpenAPI: https://fundfactsapi.com/openapi.json)
- MCP server: https://fundfactsapi.com/api/mcp (set-up for every client: https://fundfactsapi.com/mcp-server)
- Pricing: https://fundfactsapi.com/pricing
- Site index for AI assistants: https://fundfactsapi.com/llms.txt
- Informational data, not investment advice.