Quartr API

What it is

The data half of Quartr, contracted and priced separately from the Quartr Pro seat and the free mobile app. It is bought by engineering teams putting first-party company material inside something else — a research platform, a brokerage app, a newsroom tool — not by analysts.

The docs group the data into twelve datasets under a three-level hierarchy — companies own events, events contain content: companies, events, live audio, live transcripts, backlog audio, backlog transcripts, filings and reports, slide presentations, chapters, summaries, parsed documents and segments. They are bought individually or bundled, and each API key inherits its plan's permissions — a dataset outside it answers 403. The combined endpoints are all-or-nothing about that: /documents needs reports, slides and transcripts together, /live needs both live datasets.

Pricing

Nothing is published. The pricing page gives Quartr API a contact-sales button and the phrase "Enterprise & bundle options" where a number would go. No self-serve signup, no free key.

The standard terms do state the shape of the deal: a twelve-month initial period unless the order says otherwise, automatic renewal unless you give written notice 30 days before expiry, fees due 30 days from invoice, non-cancellable and non-refundable. Quartr may raise the fee at renewal on 30 days' notice, and a subscription period running past twelve months can be indexed annually — to US CPI-U in the two US contracts and to the EU-wide HICP in the two international ones — with no single adjustment exceeding eight per cent. A free evaluation period exists at Quartr's discretion, with the warranties and the indemnity switched off for it.

Data & coverage

16,000-plus companies across 65 markets and 50m-plus documents, per the vendor's own API pages. Archive depth is not published anywhere.

The docs describe the delivery SLAs as contractual: live content streamed and transcribed within five seconds of event start for 90% of events, transcripts for 95% within 45 minutes of the call ending, audio for 90% within 20 minutes, filings within 15 minutes of release, slides within 30 minutes, edited transcripts within a couple of hours. Snowflake shares refresh daily.

Audio comes as MPEG and M3U8 from a CDN with no second authentication step, backlog transcripts as JSON, reports and slides as PDFs with parsed machine-readable text behind its own licensed permission. Live transcripts stream as JSON Lines with per-word timings, confidence scores and [indiscernible] markers. Companies match on ticker, ISIN, CIK or OpenFIGI, but only the companies endpoint returns those identifiers — everything else gives you companyId and expects you to map it. Default rate limit is 50 requests a second.

Integrations

Three routes to the same data: REST at version 3, signed webhooks for created, updated and deleted records, and a Snowflake private listing a Quartr representative provisions against your account identifier. There is no SDK in any language — the docs name none — so the integration is HTTP you write yourself. Quartr MCP is a Quartr Pro feature and not part of this product.

Limitations

• Redistribution is a different contract, not a larger plan. Quartr publishes four sets of API terms — internal use and redistribution, each in a US and an international version — and the one you sign decides what the data is for.
• Under internal use, quoting is metered by word count. Derived work leaves your firm only if it is a meaningful transformation, no continuous verbatim excerpt from one document exceeds 250 words, and verbatim Quartr material does not exceed 20% of the total word count.
• Under redistribution, the data goes into one named product set out in the order and nowhere else, a Quartr logo and link must appear wherever it is displayed, and where Quartr cannot measure usage itself you owe monthly reports of what each unique user accessed, when, and whether the interaction was retrieval, summarisation or generation. Your own end users inherit the 250-word and 20% limits through Quartr's downstream data-use terms.
• Neither contract lets you train on it. No training, fine-tuning, evaluating or otherwise developing any AI or ML model on the data; an AI service may take it as input only if it does not retain the inputs or train on them. Scraping, bulk downloading, systematically caching and building a database or dataset out of the data are prohibited separately, beyond whatever the permitted purpose or the Approved Product allows.
• Nothing survives cancellation. On termination you delete, destroy or return all Quartr data, and on Quartr's request provide written certification signed by an authorised representative that you did, keeping copies only as law or a bona fide retention policy requires.
• Transcripts exist only for events conducted in English — a large hole in a 65-market product. Speaker identification is prioritised by client interest during earnings season, the name, role and company fields can be null, and older transcripts are never backfilled with it.
• No prices, estimates or financial statements. This is documents and audio about companies, not numbers out of them, and the one structured financial dataset — segments — is marked Legacy, covers only the S&P 500 and is explicitly "not offered to new partners".
• Support is your Quartr representative, 9:00–17:00 CET on weekdays, and nowhere public. The published API changelog's most recent entry is dated 18 December 2025.

Alternatives

AlphaSense sells a comparable corpus with an API and the same quote-only motion, plus broker research Quartr has none of. Fiscal.ai adds the financial statements missing here, and Daloopa is the opposite trade — numbers extracted out of the same filings rather than the documents themselves. The sibling product, Quartr, is the app a human reads; this card is the feed.

Compare it against the rest of the fundamentals and research platforms.