Toolstem Financial Data
Pricing
from $5.00 / 1,000 financial data queries
Agent-ready financial intelligence MCP tools. Stock snapshots, deep company metrics, stock screening, and side-by-side company comparison — all with derived signals and pre-computed math. Designed for AI agents, not raw API passthrough.
All notable changes to the Toolstem MCP Server will be documented in this file.
The format is based on Keep a Changelog , and this project adheres to Semantic Versioning .
screen_stocks tool temporarily disabledv1.2.1 diagnostics confirmed the root cause of screen_stocks returning 0 results: FMP's /stable/batch-quote endpoint now returns HTTP 402 ("Restricted Endpoint") on the free tier. The diagnostic payload showed every batch request rejected with a subscription-required error regardless of chunk size.
Rather than ship a broken tool, v1.2.2 removes screen_stocks from the public surface area:
src/index.ts).src/actor.ts).screen_stocks-specific fields from the Apify input schema (.actor/input_schema.json).get_stock_snapshot, get_company_metrics, compare_companies.The underlying source for the screener (src/tools/screen-stocks.ts, src/services/fmp.ts helpers, and the Russell 1000 universe in src/data/universe.ts) is retained in the repository for the v1.3 refactor, but is no longer reachable from any transport.
screen_stocks will return in v1.3 built on FMP's free-tier-available /api/v3/stock-screener endpoint (server-side filtering, 1 API call instead of 11, supports filters previously stubbed out: industry, beta, dividend, country).
_lastScreenDiag, _lastHttpStatus, _lastHttpBody, per-chunk meta.diagnostics) — retained only within the now-inert screenStocks path in fmp.ts for v1.3 reference./health endpoint and createServer() to report version 1.2.2.Post-publish QA found screen_stocks returns 0 results on the live Apify Actor despite the Russell 1000 universe being present. Apify swallows console.error output under LIMITED_PERMISSIONS, so root cause was unobservable in logs.
v1.2.1 is a diagnostic release that surfaces the failure point directly in the response body:
size, returned, ms) in meta.diagnostics.chunksmeta.diagnostics.last_http_status / last_http_bodyconsole.error to console.log so Apify captures itThis is a temporary diagnostic build — meta.diagnostics will be removed once the batch-quote issue is confirmed and fixed.
screen_stocks now works on FMP's free tierFMP's /company-screener endpoint requires a paid plan (HTTP 402 on free tier), so screen_stocks was silently returning zero results for every free-tier user. v1.2.0 replaces the call with a synthetic screener that works on the free tier:
/batch-quote endpoint and filters in-memory.Information Technology → Technology, Health Care → Healthcare).screen_stocks response: meta.universe surfaces the universe name (russell-1000), description, size (1003), and country (US). Agents can now detect scope and caveat their output appropriately.meta.unsupported_filters lists any filters that were passed but cannot be honored by the free-tier screener. meta.notes explains the limitation in plain language.console.error for visibility in Apify Actor logs.The synthetic screener cannot honor these filters on the free tier (they are accepted and surfaced as unsupported_filters in the response):
industry — iShares data has no GICS sub-industry.beta_min / beta_max — not present in /batch-quote payload.dividend_min — not present in /batch-quote payload.exchange — Russell 1000 is always NYSE/NASDAQ.country — Russell 1000 is always US (non-US country filters return unsupported).To use these filters, upgrade to an FMP Starter plan ($19/mo) and revert to the paid /company-screener endpoint.
apify was added to dependencies but left in optionalDependencies, so npm still treated it as optional and it was stripped from production builds. Removed the optionalDependencies block entirely. Apify Actor runs now start successfully.ERR_MODULE_NOT_FOUND: apify startup failure. (Superseded by 1.1.2 — this release was incomplete.)screen_stocks tool — Screen and filter stocks by sector, market cap, price, beta, volume, dividend yield, exchange, and country. Returns derived cap_category, volatility_category, and liquidity_category signals for every match. All filters optional; configurable limit up to 200.compare_companies tool — Side-by-side comparison of 2–5 companies across price, valuation (P/E, DCF), profitability (margins, ROE, ROIC), financial health, growth, dividends, and ratings. Auto-computed rankings identify leaders in each dimension: lowest_pe, highest_margin, strongest_balance_sheet, best_growth, most_undervalued, highest_rated.getBatchQuote) for efficient multi-symbol price retrieval..actor/input_schema.json updated — symbol is no longer required at the schema level (each tool validates its own inputs).MAJOR.MINOR format per platform requirements.get_stock_snapshot tool — Comprehensive stock overview combining quote, profile, DCF valuation, and rating into one response. Includes derived dcf_signal, market_cap_readable, and 52-week distance metrics.get_company_metrics tool — Deep fundamentals analysis synthesized from 5 financial statement endpoints. Includes margin_trend, health_signal, growth_signal, and pre-computed CAGRs.