MCP 服务器

Vintage

Provides point-in-time financial data access and an honest backtesting engine via MCP, enabling users to research restated fundamentals, run backtests with deflated Sharpe metrics, and benchmark returns against published factors.

README

Free financial data exists and is scattered across twenty APIs with twenty shapes. Everyone rebuilds the same glue, badly, and quietly ends up backtesting on restated figures and survivor biased universes.

Vintage is that glue, written once, served over MCP. It hosts no data. It connects, normalizes, and serves structured financial data from the web.

A research terminal that costs $0 and won't lie to you about your Sharpe.

<a href="https://rezasoleymanifar.github.io/vintage/">rezasoleymanifar.github.io/vintage</a>

What people use it for

Two of four scenes — <a href="https://rezasoleymanifar.github.io/vintage/">see the full reel on the site</a>.

Install

Claude Code

claude mcp add vintage -s user -- uvx vintage-mcp

Claude Desktop / any MCP client — add to your config file:

{
  "mcpServers": {
    "vintage": {
      "command": "uvx",
      "args": ["vintage-mcp"]
    }
  }
}

Claude Desktop config lives at %APPDATA%\Claude\claude_desktop_config.json (Windows) or ~/Library/Application Support/Claude/claude_desktop_config.json (macOS). Restart the app afterwards. MCP servers load once at startup.

Needs uv. If you'd rather use pip: pip install vintage-mcp and set the command to vintage.

Optional configuration

Everything works with zero configuration. These make it work better:

Variable	Why
`VINTAGE_USER_AGENT`	SEC EDGAR asks for a real contact. `"Your Name your@email.com"`.
`FRED_API_KEY`	Free key — unlocks 800k macro series with first-release vintages.
`VINTAGE_CACHE_DIR`	Defaults to `~/.cache/vintage`.

Set them under "env" in the same config block:

{
  "mcpServers": {
    "vintage": {
      "command": "uvx",
      "args": ["vintage-mcp"],
      "env": {
        "VINTAGE_USER_AGENT": "Jane Quant jane@example.com",
        "FRED_API_KEY": "..."
      }
    }
  }
}

Your key stays in this file. It is read by the server process and is never passed through the model or written into the conversation.

Use it as a library

The same data, without the server. Everything is synchronous and returns pandas, including inside Jupyter where a loop is already running.

import vintage as v

v.prices("AAPL", start="2020-01-01")          # daily prices, with known_at
v.panel(["AAPL", "MSFT", "JNJ"])              # dates x tickers
v.fundamentals("AAPL", "us-gaap:Assets", as_of="2020-01-01")
v.restatements("AAPL", "us-gaap:Assets")      # periods reported twice, differently
v.factors("ff3")                              # Ken French, wide
v.macro("DGS10", as_of="2008-09-15")          # ALFRED first-release vintage
v.claim("Mom12m")                             # what the paper claimed
v.claims(price_only=True)                     # the 56 replicable with free data
v.crypto("BTC-USD")
v.short_volume("AAPL")
v.sentiment("wallstreetbets")

known_at is kept as a column on every frame rather than dropped for tidiness — losing it is how a point-in-time dataset quietly becomes an ordinary one. Pass as_of and rows published after that date are gone before you see them.

Try it

Once installed, ask your assistant:

"What was Apple's total assets as of January 2020 — and has it been restated since?"

"Backtest 12-1 momentum on the Dow 30 since 2010."

"Now try short-term reversal instead. Did the alpha survive?"

The third question is the one that matters. Watch the deflated Sharpe fall as you keep asking.

The two dates

Every value carries both:

observed_at — what period the number describes
known_at — when it first became public

A backtest may only use rows whose known_at precedes the trade date. That is structural, not a setting: the panel is indexed on known_at, so any slice of it is automatically point-in-time. There is no flag to turn it off.

Sources that cannot supply an honest known_at are flagged UNKNOWN_VINTAGE rather than given a fabricated date.

Six ways yesterday's data quietly changed

Lag — the number is true in December, published in February.
Restatement — the company says "oops, wrong" and changes last year's figure.
Revision — the government keeps fixing old jobs and inflation numbers, for years.
Survivorship — dead companies get deleted; only the winners are still listed.
Membership — today's S&P 500 list is not the list from 2005.
Price adjustment — splits and dividends silently rewrite every price before them.

All six say the same thing: the data you have today is not what people saw back then.

Six verbs

Source is a parameter, never a separate tool. Twenty more sources adds zero tools.

Verb	Does
`resolve`	Any identifier → the entity key everything else accepts
`discover`	Plain-English search across every source's catalog
`fetch`	The workhorse. Any field, any source, with `as_of`
`events`	Filing timeline with exact public timestamps
`backtest`	Cross-sectional signal → returns, costs, honesty report
`benchmark`	Your returns → correlation and alpha vs published factors

Plus status for cache size, keys, and how many specs you have tried.

Backtesting methodology

Deflated Sharpe (Bailey & López de Prado, 2014) accounting for every spec tried this session

Vintage implements the backtest-validation literature rather than inventing its own statistics. Execution realism is a different problem, already solved by LEAN and Nautilus Trader — Vintage runs before that, at the stage where most ideas should die.

Technique	Source	Status
Point-in-time panel indexed on `known_at`	structural, no flag to disable	✅ shipped
Costs charged on turnover, always	no zero-cost mode exists	✅ shipped
Deflated Sharpe Ratio	Bailey & López de Prado (2014)	✅ shipped
Session trial ledger feeding the deflation	Bailey & López de Prado (2014)	✅ shipped
Probability of Backtest Overfitting, via CSCV	Bailey, Borwein, López de Prado & Zhu (2017)	⏳ planned
Purged k-fold CV with embargo	Advances in Financial Machine Learning, ch. 7	⏳ planned
Combinatorial purged cross-validation	Advances in Financial Machine Learning, ch. 12	⏳ planned
Minimum Backtest Length	Bailey, Borwein, López de Prado & Zhu (2014)	⏳ planned
Newey–West adjustment for autocorrelated returns	Newey & West (1987)	⏳ planned
Square-root market impact	Almgren et al. (2005)	⏳ planned

Citations are references, not endorsements — none of these authors is affiliated with Vintage. Anything marked planned is not in the code yet, and the backtest response says so at runtime rather than in the footnotes.

Where the data comes from

<table> <tr> <td align="center">100 years, July 1926 to this morning</td> <td align="center">331 published anomalies, with claims</td> <td align="center">18 sources, six verbs</td> <td align="center">10,398 ticker-mapped US filers</td> <td align="center">800k+ macro series with vintages</td> </tr> </table>

A century of market history, eighteen sources, and sixteen of them need no key at all. The Fama-French factors start in July 1926 and the SEC filing stream runs to this morning — Vintage covers both ends from the same six verbs.

Most of these are the primary source — not a reseller, not a scraper. The filings come from the regulator that receives them, the macro series from the central bank that publishes them, and the factors from the university that computes them.

@remove all emojis and just print list of soruceswith icons for them.

Source	Standing	Covers	Key	Point-in-time
SEC EDGAR XBRL	Primary · US regulator	Every concept every US filer has tagged, with accession number and filing date on each figure. Restatements arrive as rows, never as an overwrite.	none	✅ native filing dates
SEC filings stream	Primary · US regulator	8-K, 10-K, 10-Q, Form 4, 13D/G — timestamped to the second EDGAR accepted them.	none	✅ exact timestamps
FRED / ALFRED	Primary · central bank	Federal Reserve Bank of St. Louis. ALFRED keeps first releases, so you can ask what CPI looked like that morning.	free	✅ first-release vintages
Ken French Data Library	Primary · academic	Dartmouth. FF3, FF5, momentum, daily FF3, 49 industry portfolios — from where the authors publish them.	none	❌ rebuilt each release
Open Source Asset Pricing	Primary · academic	Chen & Zimmermann. 331 published predictors with claimed return, t-stat, sample window and an implementable definition. `openap:Mom12m` returns Jegadeesh-Titman's 1.31%/mo, t=3.74.	none	✅ claims dated to publication year
SEC Form 13F	Primary · US regulator	Institutional equity holdings for every manager over $100m. Quarter end and filing date are up to 45 days apart and both are kept, so `as_of` returns the book that was actually public.	none	✅ quarter end vs filing date
SEC Form 25	Primary · US regulator	Every delisting on record — 36,830 filings across 11,614 companies. The correction for a universe built from names that still exist.	none	✅ filing dates, never revised
SEC XBRL frames	Primary · US regulator	One concept across every filer in a single call. 6,289 companies in 840 KB — the shape a cross-sectional sort needs.	none	❌ carries the accession, not its date
US Treasury	Primary · US government	The par yield curve, 14 tenors from one month to thirty years, published each business day.	none	✅ never revised
CFTC	Primary · US regulator	Commitments of Traders. Tuesday's positioning by trader class, released the following Friday, and the lag is preserved.	none	✅ lag preserved in `known_at`
Bureau of Labor Statistics	Primary · US agency	CPI down to item strata, payrolls, JOLTS, wages, productivity. Any series id, not a curated shortlist.	optional	❌ ships no release date
Bureau of Economic Analysis	Primary · US agency	The national accounts. One call returns every line of a NIPA table rather than one series at a time.	free	❌ current estimate only
European Central Bank	Primary · central bank	Daily FX reference rates since 1999, plus any cross derived from two euro legs and labelled as derived.	none	✅ published once, never revised
CBOE	Primary · exchange	VIX and the whole volatility family — term structure, VVIX, SKEW — back to 1990.	none	✅ index levels are not revised
FINRA	Primary · US regulator	Daily short sale volume per symbol, published after each close and never revised. Short volume, not short interest.	none	✅ never revised
Coinbase Exchange	Exchange	Crypto OHLCV, every listed pair.	none	✅ trade prints are never restated
ApeWisdom	Community	Forum mention ranks across ~15 subreddits. No history upstream — rows are stamped when Vintage fetched them.	none	⏩ forward only, from the day you record
Yahoo Finance	Third party	Daily OHLCV and adjusted close, decades deep.	none	⚠️ adjusted retroactively, flagged on every row

COVERAGE.md is the full field-by-field catalogue — every prefix, every dataset, every signal, with measured coverage spans. It is generated from the registry, so it cannot drift from the code.

Counts current as of August 2026. Vintage redistributes none of this — each upstream source keeps its own terms.

Cache

Gzipped JSON in ~/.cache/vintage, tiered by how mutable the data is: closed periods never refetch, academic datasets monthly, current fundamentals daily, prices per session. An hour of conversation is roughly 20 upstream calls.

Known gaps

Stated plainly, because the alternative is shipping a bad substitute:

Data:

Survivorship — universes are current-listing only. Form 25 delistings are the next build and the backtester warns until then.
Analyst estimates — no free source exists.
Historical options chains — paid everywhere.
Point-in-time index membership — licensed by S&P and MSCI.

Engine — the backtester is vectorized and cross-sectional, which is a rung below an event-driven simulator:

No purging or embargo — overlapping label windows can leak across a train/test split (López de Prado, AFML ch. 7). Deflation catches selection bias, not leakage.
No market impact — costs are a flat charge on turnover, so large-notional results are optimistic.
No PBO — deflated Sharpe covers multiple testing; the Probability of Backtest Overfitting via combinatorially symmetric cross-validation would be the stronger test.
Trial count resets each session — ask forty things today and forty tomorrow, and tomorrow starts from zero.
Sharpe is per observation, not annualized — that is the frequency the deflation is defined at, and the response says so.

Development

git clone https://github.com/RezaSoleymanifar/vintage
cd vintage
uv sync --group dev
uv run pytest

smoke_test.py exercises all six verbs against the live sources — useful before a release, and it needs network.

License

MIT. Vintage redistributes no data; each upstream source keeps its own terms.

mcp-name: io.github.RezaSoleymanifar/vintage