The problem
Speed is table stakes. Being right is the hard part.
Frontier models are brilliant at language and reasoning. But three structural gaps — that no amount of model scaling fixes — mean an AI alone can't be trusted with serious research.
It doesn't know what happened after training
The latest 10-K, this morning's 13D, last quarter's 13F — none of it exists for a model frozen at its training date. With MCP, your AI queries filings, news and holdings as they're published — and cites them.
It can't reconstruct "what was true on date X"
Ask for a company as of Q3 2018 and you get today's picture (wrong) or a confident reconstruction (untrustworthy). Every Kaleidoscope tool accepts an "as of" date and returns what was knowable then — leak-safe by construction.
Its base rates are the winners' success rate
Training corpora are dominated by the companies and drugs that got written about. Kaleidoscope computes base rates over survivorship-complete populations — failures, delistings and discontinued programs in the denominator.
In practice
What one question gets you
Each of these is a single natural-language question. Your AI composes the tools and returns a synthesized read — every fact traceable to its source.
"Brief me on this name."
A synthesized read — signals, 8-K activity, insider trades, latest quarter — each line cited to the filing that established it. An hour of tab-switching, compressed to one answer.
get_company_profile get_fundamentals
"Who's been quietly accumulating?"
Cluster entries, quiet accumulation and consecutive-quarter adds screened across hundreds of notable funds — share-based actual trading, not price-inflated values.
thirteenf_screen thirteenf_smart_money_consensus
"Watch my list — anything moving?"
A whole watchlist scanned against pre-aggregated signals over 152M filing sentences and 127M news sentences — a multi-hour batch job, returned sub-second.
screen_companies search_classifiers
The difference
Why this is hard to replicate
"Couldn't someone just point an LLM at SEC EDGAR?" The short answer is no. The long answer is years of work:
Extraction pipelines tuned over years
The 529-classifier taxonomy and 152M classified sentences aren't a corpus you can spin up — they're the output of an extraction system tuned against real outcomes.
The survivorship-complete denominator
Including delistings and failed programs means reconstructing companies that no longer exist from filings no aggregator surfaces. The base rate is the moat.
The provenance contract
Every fact carries its source, snippet and as-of date. Most APIs return data; this returns data plus verifiability — which is what makes an AI agent trustworthy inside it.
Continuous freshness
Filings flow through as the SEC publishes them, 13F refreshes every ~4 hours, news daily. A competitor would need to match not just the data but the cadence.
The bottom line
A faster wrong answer is still wrong
The hard part was never the chat — it's whether the answer is true as of the date you asked, whether the base rate counts the failures, and whether every claim traces to a source you can open. That's the layer Kaleidoscope MCP adds.
FAQ
Common questions
What is the Model Context Protocol (MCP)?
MCP is an open standard that lets AI assistants like Claude, ChatGPT, Gemini, and Cursor call external tools and data sources directly. Kaleidoscope MCP is a family of MCP servers that connect those assistants to cited SEC, SEDAR, and market research data.
What is Kaleidoscope MCP?
Kaleidoscope MCP is a family of read-only research servers — one per domain — that give any MCP-compatible AI client direct access to SEC filings, institutional 13F flow, activist 13D/13G campaigns, fundamentals, news, Canadian SEDAR filings, and more — every answer cited to a source filing and accurate as of any historical date.
Which AI clients does Kaleidoscope MCP work with?
Any client that supports the Model Context Protocol over HTTP — including Claude (Desktop, Code, and claude.ai), ChatGPT, Google Gemini, Cursor, and custom agents.
How do I connect Kaleidoscope MCP to Claude or ChatGPT?
Request a demo and we'll allowlist your email for the servers your engagement covers. Add a server's endpoint URL to your client's MCP settings; on first use your client opens a one-time email sign-in — no API keys or bearer tokens to manage. Every tool is read-only and side-effect-free, so clients can auto-approve them.
How do I get access to Kaleidoscope MCP?
Access is by invitation, per server. Request a demo and we'll walk through the data on your own questions, then allowlist your team's emails for the servers you need.
What data can I access through Kaleidoscope MCP?
Ten live servers across the market: SEC filings, 13F flow and activist ownership, XBRL fundamentals, news, Canadian SEDAR and NI 43-101 technical reports, private-company funding, material agreements, M&A deals, biotech pipelines, catalysts, power, and corporate debt. Every fact returns its source filing and as-of date.