Connect your AI assistant
https://mcp.csimarket.com/mcp
- Settings → Connectors → Add custom connector
- Name it CSIMarket, paste the URL above
- Add, then Connect, then sign in
- Settings → Apps & Connectors → Developer mode
- Create a connector with the URL above
- Authentication: OAuth, then sign in
- Settings → Tools & MCP → Add custom MCP
- Point the server url at the URL above
- Click Connect and sign in
- Standards-compliant: Streamable HTTP + OAuth 2.1 discovery
- Point it at the URL above
- Let it complete OAuth discovery + sign-in
Turn your AI into a
financial research analyst
Connect ChatGPT, Claude, and other MCP-compatible AI assistants to CSIMarket's structured financial intelligence. Research companies, compare competitors, analyze industries, screen thousands of companies, investigate supply chains, and work with historical and quantitative financial data, using live CSIMarket data instead of static AI knowledge.
Net margin 18.4% · Revenue growth +12.1%
P/E median 24.8× · ROE 22.3%
vs. all-industry median: Net margin 8.1%
Semiconductors rank 3rd in profitability across 110 industries.
What you can ask
The point is not any single number. It is that one conversation can cross datasets that normally live in separate products: compare semiconductor competitors, examine their segment exposure, trace supplier relationships, check historical fundamentals as they were actually reported, and weigh the quantitative factors, without leaving the chat.
One prompt, an institutional-style profile: fundamentals, competitors, segment and geographic exposure, and the supply-chain relationships behind the business.
Put a company beside the competitors its own filings name, on the measures that decide who is winning.
Size a company against the competitors its filings name, and see who leads the set.
Compare margins, growth rates, and valuation multiples across industries. Identify the best-performing sectors and understand why.
Pull a complete financial snapshot fundamentals, valuation, efficiency and benchmark it against the company's industry peers.
Screen thousands of companies in one prompt. Filter by sector, growth rate, margin thresholds, or quant factor scores.
Read the market as a whole: which regime it is in, how confident that call is, and which sectors and industries are gaining or losing leadership.
Ask what was actually known on a past date. Figures come back as first reported, as later revised, or as they stood on your as-of date, so a historical study is not quietly answered with today's numbers.
Trace who a company buys from and sells to, then ask for the filings behind each link rather than taking the graph on trust.
Identify a company's competitors, compare them on revenue, growth and margins, and work out who leads the market and who is gaining.
Follow the relationship graph past the first hop: how two companies connect, what sits in a company's wider neighbourhood, and where indirect exposure concentrates by industry.
Financial data, delivered to your AI
The CSIMarket MCP Server implements the Model Context Protocol an open standard that lets AI assistants securely access external tools and data sources. Instead of copying data into a prompt, your AI calls CSIMarket directly, gets live results, and reasons over them always fresh, always cited.
Works with Claude (Anthropic), ChatGPT (OpenAI), and any application built on the MCP standard. Connect with your CSIMarket subscription MCP key; your plan's data scope is enforced on every request.
# 1. Add the MCP server to your AI client url: https://mcp.csimarket.com/mcp # 2. Authenticate with your CSIMarket subscription MCP key # (add it to your AI client's MCP config see below) # 3. Ask questions naturally You: "Compare pharma vs biotech margin trends" # 4. AI calls CSIMarket tools with your live data → csimarket_industry(id=73, period="TTM") → csimarket_industry(id=12, period="TTM") # 5. AI synthesizes a structured answer from live data
Built for serious financial analysis
Live data, every query
No static embeddings or stale context. Every tool call hits the live CSIMarket database TTM figures updated daily, fundamentals refreshed within hours of SEC filing ingestion.
Your plan, enforced per call
Subscription key authentication. Each tool call is validated against your CSIMarket license scope restrictions and rate limits apply automatically to every request.
Institutional-depth data
110 industries · 13 sectors · 13,000+ companies. Industry benchmarks with statistical distributions, quantitative factor panels with 90 fields, and 20+ years of normalized fundamentals not a summary layer.
Native multi-client support
The same MCP server works with Claude Desktop, Claude.ai, ChatGPT, and any application that supports the MCP protocol one integration, every client.
Cross-industry screening
Screen thousands of companies across fundamentals, valuation, and quant factors in a single tool call. The AI can filter, rank, and explain results in conversational output.
Source-cited answers
Responses come from structured CSIMarket data, not hallucinated text. The AI returns the raw numbers alongside its synthesis, so you can verify every figure independently.
Who leads a market, and who is catching up
Identify competitors, compare financial performance, measure market share, and track how competitive position changes over time.
Compare revenue, growth, profitability and valuation across a company and the competitors its filings name.
Measure a company's revenue against the combined revenue of its competitors, with rank and coverage.
Compare business segments where company reporting definitions can be reliably aligned, against a consolidated denominator that is published with every response.
Analyze disclosed suppliers, customers and multi-hop relationships, with the filing behind each link available on request.
Track how competitive position moves across reporting periods, using the figures as they were reported at the time.
Compare where revenue actually comes from by region and country, as independent partitions that are never summed across.
61 tools, 6 data domains
Each tool maps to a CSIMarket API endpoint. The AI picks the right tool automatically
from your question, so there is nothing to configure. Scope badges show what gates a
tool; csimarket_my_access reports exactly what your own licence allows.
-
csimarket_coverage
What data CSIMarket holds: live entity counts, available datasets, period ranges and methodology.
All plans -
csimarket_connection_status
Confirms the MCP connection itself is live, distinct from what it is licensed to read. The authoritative answer to "are you connected to CSIMarket?"
All plans -
csimarket_my_access
What this connection is licensed to read, and how much of its rate limit is left. Use it to settle any "should I be able to see this?" question.
All plans -
csimarket_market_overview
Total US market aggregates: the whole listed universe treated as one entity.
Plan scoped -
csimarket_industry
Industry overview and fundamental metrics across margins, growth, valuation, financial strength and efficiency. Supports TTM, FY and quarterly.
industries.read -
csimarket_sector
Sector overview and metrics, the same measure set rolled up to 13 sectors.
industries.read -
csimarket_rankings
Rank all industries or sectors by a chosen metric.
Plan scoped -
csimarket_industry_search
Find the right industry or sector from a name or rough description. Tolerates typos and colloquial names.
Plan scoped -
csimarket_search
Resolve a company by name or ticker, or find industries and sectors by name.
Plan scoped -
csimarket_company_search
Find a company by name when you have no ticker or CIK. Typos, partial names and spacing variants all resolve.
Plan scoped -
csimarket_company_overview
Company identity and headline metrics. The right first call for almost any company question.
companies.read -
csimarket_company_financials
Financial statements and ratio categories for a company, returning only the sections you request.
companies.read -
csimarket_company_history
Historical time series for a company metric or financial statement.
statements.read -
csimarket_peer_analysis
Compare a company against its industry benchmark.
companies.read -
csimarket_compare
Compare multiple companies side by side across a chosen dimension.
Plan scoped -
csimarket_company_ownership
Institutional ownership of a company, from 13F filings.
Plan scoped -
csimarket_institutional_holder
The manager side of 13F: what one institutional investor holds.
Plan scoped -
csimarket_news
CSIMarket news and corporate-event feed.
Plan scoped -
csimarket_news_story
Read one news item in full, either a single article by id or an entire story cluster.
Plan scoped -
csimarket_company_quant
The institutional differentiator: 90-field quant factor panel, betas, risk scores, z-scores and percentile ranks on a point-in-time basis.
quant.read -
csimarket_screen
Screen the quant-factor universe of roughly 10,020 companies using quantitative factor filters.
quant.read -
csimarket_company_relationships
The commercial relationship graph for a company: who it buys from, sells to, and competes with.
relationships.read -
csimarket_company_relationship_history
Batch relationship history for an entire company in one call.
relationships.read -
csimarket_relationship_history
Point-in-time history of a single relationship: the observable timeline behind it.
relationships.read -
csimarket_relationship_evidence
Provenance for a single relationship. Answers "why do you believe this?" with the underlying source.
relationships.evidence.read -
csimarket_relationship_values
Disclosed economic value of a single relationship as a point-in-time observation series.
relationships.value.read -
csimarket_relationship_analytics
Derived supply-chain and counterparty-risk datasets for one company.
Plan scoped -
csimarket_relationships_compare
Compare the counterparties of several companies in one call: which are shared and which are unique.
Plan scoped -
csimarket_supply_chain_exposure
Concentration and dependency across a company's disclosed customers and suppliers, with the evidence basis labelled rather than assumed.
relationships.read -
csimarket_graph_neighborhood
The multi-hop neighbourhood around a company: who its counterparties trade with, returned as nodes and edges rather than a flat list.
relationships.read -
csimarket_graph_paths
How two companies are connected. Returns every disclosed route between them, so an indirect dependency two or three hops away becomes visible.
relationships.read -
csimarket_graph_exposure
Indirect exposure across the whole graph, with an industry rollup and weighted propagation, plus the coverage behind it.
relationships.read -
csimarket_market_share
Whole-company revenue market share vs named competitors, across the full peer set, industry and sector, with each company's segment breakdown attached.
relationships.read -
csimarket_market_share_history
The same market share as an annual TTM series, so a gain or loss of position reads as a trend across reporting periods rather than a single reading.
relationships.read -
csimarket_company_segments
Reported business segments for a company: revenue and operating income per member against the published consolidated denominator.
segments.read -
csimarket_company_geographic
Reported geographic revenue exposure, served as region and country views that are independent partitions, never summed across.
segments.read -
csimarket_company_exposure
Business-segment and geographic revenue exposure merged into one call, the combined view when you need both dimensions together.
segments.read -
csimarket_segment_detail
Drill into one specific reported segment or geographic member across time, by its own id.
segments.detail.read -
csimarket_segment_screener
Screen the whole company universe by segment or geographic revenue exposure, e.g. every company with 20%+ revenue from one region.
segments.screener.read -
csimarket_insider_activity
SEC Forms 3/4/5 insider trading: officer, director and 10% owner transactions, plus a net buy/sell summary.
ownership.read -
csimarket_ownership_screener
Screen the whole company universe by institutional-ownership concentration and composition.
ownership.read -
csimarket_stock_price
Share price for the latest session or any past trading day: open/high/low/close, previous close, change, returns, volume and ATR.
stocks.read -
csimarket_stock_price_history
Daily price history back to 2006, newest first, across any date range.
stocks.read -
csimarket_market_performance
Cap-weighted performance of the whole listed universe with price breadth, over windows from one day to a full year.
market.read -
csimarket_market_movers
The session's gainers, losers and most-active names, for today or for any past trading day.
market.read -
csimarket_index_composite
Benchmark quotes for the major US indices, with constituent aggregates for the current session.
market.read -
csimarket_market_regime
Which regime the market is in, from a rules-based vote over return windows, breadth and volatility. The formula is returned with the answer and the confidence is a deterministic agreement share, never a black-box probability.
market.read -
csimarket_sector_rotation
Which sectors are leading, lagging, improving or deteriorating, with rank changes measured against an earlier session.
market.read -
csimarket_industry_rotation
The same rotation picture at industry level: momentum, rank change and rotation candidates.
market.read -
csimarket_sector_performance
Cap-weighted sector league table across trading-day windows.
market.read -
csimarket_industry_performance
Cap-weighted industry league table across the same windows.
market.read -
csimarket_research_company
A complete company dossier in one call: up to 13 sections composed from the underlying data, each section gated by your own licence so you get everything you are entitled to and nothing you are not.
companies.read -
csimarket_pit_fundamentals
What was actually known on a past date. Ask for a figure as first reported, as later revised, or as it stood on any as-of date, without today's hindsight leaking into the answer.
Plan scoped -
csimarket_earnings_dates
When results were actually announced. For each 10-K and 10-Q, the earnings announcement date alongside the date the formal statement was filed, with the gap between them. About 28% of reports have no separate announcement, and those are returned as such rather than guessed.
fundamentals.read -
csimarket_analysis_guide
The AI's own map of the data: which tool answers which question, and the conventions behind the numbers.
All plans -
csimarket_export_report
Turn an analysis into a structured report file for downstream work.
Plan scoped -
csimarket_company_changes
What changed at a company between periods, across fundamentals and reported data.
Plan scoped -
csimarket_company_discover
Find companies from a description when you have no name or ticker to start from.
Plan scoped -
csimarket_company_similar
The companies most similar to a given one: peer sets that are not merely an industry code.
Plan scoped -
csimarket_event_impact
How a company's figures moved around a corporate event.
Plan scoped -
csimarket_portfolio_exposure
Aggregate exposure across a list of holdings: where a portfolio's industry and counterparty concentrations actually sit.
Plan scoped
-
csimarket_coverage
What data CSIMarket holds: live entity counts, available datasets, period ranges and methodology.
All plans -
csimarket_connection_status
Confirms the MCP connection itself is live, distinct from what it is licensed to read. The authoritative answer to "are you connected to CSIMarket?"
All plans -
csimarket_my_access
What this connection is licensed to read, and how much of its rate limit is left. Use it to settle any "should I be able to see this?" question.
All plans -
csimarket_market_overview
Total US market aggregates: the whole listed universe treated as one entity.
Plan scoped -
csimarket_industry
Industry overview and fundamental metrics across margins, growth, valuation, financial strength and efficiency. Supports TTM, FY and quarterly.
industries.read -
csimarket_sector
Sector overview and metrics, the same measure set rolled up to 13 sectors.
industries.read -
csimarket_rankings
Rank all industries or sectors by a chosen metric.
Plan scoped -
csimarket_industry_search
Find the right industry or sector from a name or rough description. Tolerates typos and colloquial names.
Plan scoped -
csimarket_search
Resolve a company by name or ticker, or find industries and sectors by name.
Plan scoped -
csimarket_market_regime
Which regime the market is in, from a rules-based vote over return windows, breadth and volatility. The formula is returned with the answer and the confidence is a deterministic agreement share, never a black-box probability.
market.read -
csimarket_sector_rotation
Which sectors are leading, lagging, improving or deteriorating, with rank changes measured against an earlier session.
market.read -
csimarket_industry_rotation
The same rotation picture at industry level: momentum, rank change and rotation candidates.
market.read -
csimarket_sector_performance
Cap-weighted sector league table across trading-day windows.
market.read -
csimarket_industry_performance
Cap-weighted industry league table across the same windows.
market.read
-
csimarket_company_search
Find a company by name when you have no ticker or CIK. Typos, partial names and spacing variants all resolve.
Plan scoped -
csimarket_company_overview
Company identity and headline metrics. The right first call for almost any company question.
companies.read -
csimarket_company_financials
Financial statements and ratio categories for a company, returning only the sections you request.
companies.read -
csimarket_company_history
Historical time series for a company metric or financial statement.
statements.read -
csimarket_peer_analysis
Compare a company against its industry benchmark.
companies.read -
csimarket_compare
Compare multiple companies side by side across a chosen dimension.
Plan scoped -
csimarket_company_ownership
Institutional ownership of a company, from 13F filings.
Plan scoped -
csimarket_institutional_holder
The manager side of 13F: what one institutional investor holds.
Plan scoped -
csimarket_news
CSIMarket news and corporate-event feed.
Plan scoped -
csimarket_news_story
Read one news item in full, either a single article by id or an entire story cluster.
Plan scoped -
csimarket_company_segments
Reported business segments for a company: revenue and operating income per member against the published consolidated denominator.
segments.read -
csimarket_company_geographic
Reported geographic revenue exposure, served as region and country views that are independent partitions, never summed across.
segments.read -
csimarket_company_exposure
Business-segment and geographic revenue exposure merged into one call, the combined view when you need both dimensions together.
segments.read -
csimarket_segment_detail
Drill into one specific reported segment or geographic member across time, by its own id.
segments.detail.read -
csimarket_segment_screener
Screen the whole company universe by segment or geographic revenue exposure, e.g. every company with 20%+ revenue from one region.
segments.screener.read -
csimarket_insider_activity
SEC Forms 3/4/5 insider trading: officer, director and 10% owner transactions, plus a net buy/sell summary.
ownership.read -
csimarket_ownership_screener
Screen the whole company universe by institutional-ownership concentration and composition.
ownership.read -
csimarket_company_changes
What changed at a company between periods, across fundamentals and reported data.
Plan scoped -
csimarket_company_discover
Find companies from a description when you have no name or ticker to start from.
Plan scoped -
csimarket_company_similar
The companies most similar to a given one: peer sets that are not merely an industry code.
Plan scoped -
csimarket_event_impact
How a company's figures moved around a corporate event.
Plan scoped -
csimarket_portfolio_exposure
Aggregate exposure across a list of holdings: where a portfolio's industry and counterparty concentrations actually sit.
Plan scoped
-
csimarket_company_quant
The institutional differentiator: 90-field quant factor panel, betas, risk scores, z-scores and percentile ranks on a point-in-time basis.
quant.read -
csimarket_screen
Screen the quant-factor universe of roughly 10,020 companies using quantitative factor filters.
quant.read
-
csimarket_company_relationships
The commercial relationship graph for a company: who it buys from, sells to, and competes with.
relationships.read -
csimarket_company_relationship_history
Batch relationship history for an entire company in one call.
relationships.read -
csimarket_relationship_history
Point-in-time history of a single relationship: the observable timeline behind it.
relationships.read -
csimarket_relationship_evidence
Provenance for a single relationship. Answers "why do you believe this?" with the underlying source.
relationships.evidence.read -
csimarket_relationship_values
Disclosed economic value of a single relationship as a point-in-time observation series.
relationships.value.read -
csimarket_relationship_analytics
Derived supply-chain and counterparty-risk datasets for one company.
Plan scoped -
csimarket_relationships_compare
Compare the counterparties of several companies in one call: which are shared and which are unique.
Plan scoped -
csimarket_supply_chain_exposure
Concentration and dependency across a company's disclosed customers and suppliers, with the evidence basis labelled rather than assumed.
relationships.read -
csimarket_graph_neighborhood
The multi-hop neighbourhood around a company: who its counterparties trade with, returned as nodes and edges rather than a flat list.
relationships.read -
csimarket_graph_paths
How two companies are connected. Returns every disclosed route between them, so an indirect dependency two or three hops away becomes visible.
relationships.read -
csimarket_graph_exposure
Indirect exposure across the whole graph, with an industry rollup and weighted propagation, plus the coverage behind it.
relationships.read -
csimarket_market_share
Whole-company revenue market share vs named competitors, across the full peer set, industry and sector, with each company's segment breakdown attached.
relationships.read -
csimarket_market_share_history
The same market share as an annual TTM series, so a gain or loss of position reads as a trend across reporting periods rather than a single reading.
relationships.read
-
csimarket_stock_price
Share price for the latest session or any past trading day: open/high/low/close, previous close, change, returns, volume and ATR.
stocks.read -
csimarket_stock_price_history
Daily price history back to 2006, newest first, across any date range.
stocks.read -
csimarket_market_performance
Cap-weighted performance of the whole listed universe with price breadth, over windows from one day to a full year.
market.read -
csimarket_market_movers
The session's gainers, losers and most-active names, for today or for any past trading day.
market.read -
csimarket_index_composite
Benchmark quotes for the major US indices, with constituent aggregates for the current session.
market.read
-
csimarket_research_company
A complete company dossier in one call: up to 13 sections composed from the underlying data, each section gated by your own licence so you get everything you are entitled to and nothing you are not.
companies.read -
csimarket_pit_fundamentals
What was actually known on a past date. Ask for a figure as first reported, as later revised, or as it stood on any as-of date, without today's hindsight leaking into the answer.
Plan scoped -
csimarket_earnings_dates
When results were actually announced. For each 10-K and 10-Q, the earnings announcement date alongside the date the formal statement was filed, with the gap between them. About 28% of reports have no separate announcement, and those are returned as such rather than guessed.
fundamentals.read -
csimarket_analysis_guide
The AI's own map of the data: which tool answers which question, and the conventions behind the numbers.
All plans -
csimarket_export_report
Turn an analysis into a structured report file for downstream work.
Plan scoped
Up and running in minutes
No SDKs to install. Add the server URL and your subscription MCP key to your AI client's config your assistant handles the rest.
Register at csimarket.com and choose a plan. Your plan determines which tools and data your AI can access.
In Claude Desktop, ChatGPT, or your MCP client, add the CSIMarket server URL and your subscription MCP key. Find your key in the API dashboard.
No special syntax needed. Ask about industries, companies, or markets in plain language your AI picks the right tools automatically.
Configuration
// ~/Library/Application Support/Claude/claude_desktop_config.json { "mcpServers": { "csimarket": { "url": "https://mcp.csimarket.com/mcp" } } }
No key goes in the file. On first use the client opens CSIMarket sign-in and stores a
short-lived token itself. Some older docs show an Authorization: Bearer
header instead; that works, but the token expires within the hour, so let the client
sign in. Either way the credential lives in the config, never in a prompt.
{
"mcpServers": {
"csimarket": {
"transport": "http",
"url": "https://mcp.csimarket.com/mcp"
}
}
}
# Server metadata https://mcp.csimarket.com/.well-known/oauth-authorization-server # Health & version https://mcp.csimarket.com/health https://mcp.csimarket.com/version
Enterprise-grade access control
Your MCP key stays in your AI client's config it never appears in prompts or conversation. Every data access is scoped to your subscription plan and enforced per request.
Subscription key auth
Your MCP key from your CSIMarket subscription. Set it once in your AI client's config never paste it into a prompt.
Per-request scope enforcement
Each tool call is validated against your plan's entitlements. Quant factor tools require an explicit quant.read scope.
License-level enforcement
Every request is validated against your live license status. A suspended or expired subscription stops access immediately no waiting for tokens to expire.
TLS-only, no plaintext
All traffic between your AI and the MCP server is encrypted over HTTPS. Data is never stored in the AI's conversation.
Excel and Power BI via OData 4.0
For spreadsheet and BI workflows, CSIMarket also offers an OData 4.0 connector connect Excel or Power BI directly to CSIMarket data without writing any code. 41 entity sets mirror the full API surface.
# Data → Get Data → From OData Feed https://api.csimarket.com/api/odata/v1 # Free sample (no key, industry 201 data) https://api.csimarket.com/api/odata/test
LLM sitemaps
Machine-readable maps of what CSIMarket covers, written for language models rather than for search-engine crawlers. An agent that reads these knows which datasets exist, which MCP tool answers which question, and how each tool maps to a REST endpoint and a page on the site, before it makes a single call.
# Concise index https://csimarket.com/llms.txt # Full map: tools, endpoints, site sections https://csimarket.com/llms-full.txt
Start using CSIMarket in your AI today
Choose the plan that fits your work and connect the MCP server to Claude or ChatGPT in minutes.