AlphaSense

What it is

AlphaSense is a search engine over documents an institution would otherwise read one at a time — SEC and global filings, earnings transcripts, news, regulatory announcements, sell-side and independent equity research, and the Tegus library of expert-call transcripts. The vendor puts the corpus at over 500 million documents. The corpus is what the contract is actually for.

The layer it leads with in 2026 is AI — Generative Search, an agent called SuperAnalyst, and Deep Research, which writes multi-page briefings. Answers carry inline citations back to the source document, returned over the API as Markdown with [[N • Source Name]] markers. The homepage claims outputs "with sentence-level citations and no hallucinations"; treat the first half as a checkable feature and the second as marketing.

Several familiar names are now this one product. Sentieo was acquired in May 2022 and Tegus in July 2024, bringing BamSEC and the Canalyst model library with it.

Pricing

Quote-only, and the pricing page is unusual in publishing no number whatsoever. Two plans: Market Intelligence, and Enterprise Intelligence, which adds federated search over your own internal documents plus private-cloud and bring-your-own-bucket hosting. The vendor's wording is annual subscriptions "for all team sizes, ranging from enterprise packages to per-seat options" — so the unit is a negotiated seat on a year's contract.

Two things sit outside both plans as paid add-ons. Expert calls — arranging a live call with an operator from the one-million-expert network — are separate from the transcript library, which is listed as included content in both plans. Canalyst financial models are likewise an add-on.

No free tier, and no self-serve trial — a trial exists but is arranged by sales, with no published length.

Data & coverage

Global, and weighted toward what an analyst reads rather than what a chart plots. Broker and independent research spans 1,000+ sell-side and independent firms; the Tegus library is 300,000+ expert interviews across roughly 29,000 public and private companies, growing by about 8,000 a month. Financial data covers 18,000+ public companies with standardised statements and forward consensus, plus 1.4 million private companies and a screener the vendor puts at 400+ filters. Market data is thin — "stock pricing and volume metrics" on a company profile, not a feed. Archive depth is published for no content set, which matters if you are buying this to read something from 2006.

Integrations

A GraphQL API covers search, companies, brokers, watchlists and document download; a separate Agent API for Generative Search is metered in AI credits — 10 per fast query, 25 for ThinkLonger, 100 for a Deep Research run, against a limit set in your commercial agreement. A first-party MCP connector exposes two tools, alphasense_search and alphasense_search_specific, but the account team must enable it and an administrator must add it for the organisation. PowerPoint and Excel add-ins, iOS and Android apps, and SSO round it out. API access is not automatic either — the docs tell you to ask your account team.

Limitations

• You almost certainly cannot buy this. No self-serve tier, no published price, annual contracts, seat-licensed, sold to firms. A retail reader's nearest route to the same corpus is the same company's self-serve filings product, BamSEC.
• Real-time broker research is not included by virtue of paying AlphaSense. It needs a per-broker entitlement, and brokers grant those to clients with a trading, advisory or coverage relationship. Without one you read aftermarket research on a 0–7 day embargo.
• That aftermarket research is metered by the page. Opening keyword hits, scrolling between pages and re-opening a page more than 24 hours later all draw down an allowance your account manager sets.
• Expert calls and Canalyst models are priced separately from the platform.
• API and MCP access are gated behind the account team, and AI usage is credit-limited per contract rather than flat.
• It is not a market-data or portfolio product — no charting, no positions, no execution.

Alternatives

Against the terminals — Bloomberg, S&P Capital IQ, FactSet, LSEG Workspace — it competes on the document corpus and the search over it, not on pricing data. Against GLG and Third Bridge it competes with a transcript library it owns outright. If what you wanted was filings search rather than an enterprise contract, the cheap end of that job is in fundamentals and research platforms, and the AI layer has narrower, far cheaper equivalents in MCP servers and AI agents.