FinTech Studios

The removal test leaves a news archive. Natural language processing operates across more than 100 languages on billions of signals, generative search returns conversational answers with contextually suggested prompts, and separate products handle regulatory tracking, market analysis and chart explanation.

The company describes the platform as a continuously running intelligence substrate rather than a database, and its newest positioning treats artificial intelligence agents as a first class user of the engine alongside people.

Output is research rather than decision, delivered with a property the company names explicitly: insights are attributable and fact based, so a user can follow any generated statement back to the source that supports it, which is the correct control for a research product where the failure mode is a confident fabrication reaching an investment or compliance judgement. Suggested prompts guide rather than automate.

The qualification is the newest direction, since the platform is now positioned for consumption by autonomous agents, and nothing describes what verification applies when a machine rather than an analyst is the reader.

Attribution is the stated design property and the right one, with generative output described as real time, attributable and fact based, grounded in millions of authoritative sources so that any claim can be checked against what produced it. Naming the underlying model provider adds a second layer of transparency, since a buyer can reason about the known behaviour of that model rather than treating the generation step as opaque.

What is missing is measurement: no accuracy, hallucination rate or coverage completeness figure is published, and for a regulatory intelligence product the rate of missed rule changes is the number that matters most.

Twelve years of operation with no customer named anywhere is the central gap, and the claim that the engine is trusted by the world's top financial institutions is unattributed. Headcount stands around eleven across two continents, and total funding is modest at roughly six million dollars with the last disclosed round in 2017 or 2019 depending on the source, which means the business has run on revenue for the better part of a decade.

That is either resilience or constraint and nothing published distinguishes them. Supporting evidence is thin but real: a named integration partner announced in March 2026 for white labelled video briefings, and a software marketplace leader badge for autumn 2025.

No data boundary statement was located, and the enterprise search capability makes the question concrete. The platform is described as seamlessly integrating millions of authoritative sources of web and enterprise content, which means a customer's own internal documents are indexed alongside public material, and nothing states whether that enterprise content is isolated per customer, whether queries are retained, or how the newer self serve tier is separated from enterprise deployments.

Structurally favourable because the corpus is public and market facing: news, regulatory filings, market data and social media, augmented by premium and private licensed sources, with no consumer records processed and no individual the subject of an assessment. Social media ingestion does mean personal content enters the index, though as market signal rather than as a profile of any person. Held at B because the platform also integrates enterprise content belonging to customers, and no data processing terms, retention schedule or subprocessor list was located.

An enterprise security badge appears on a software marketplace listing and no attestation, certification, trust centre or enumerated framework was located. For a platform indexing customers' internal enterprise content alongside public sources, and now offering a free self serve tier on the same engine, published assurance about how those two are separated would be the natural disclosure and none exists.

No supervisor, statute or instrument governing the company itself is named. One product tracks regulatory change for customers, which means the platform reads the rulebook rather than being bound by it, and that distinction is worth keeping since a regulatory intelligence tool whose coverage is incomplete creates compliance risk for the institution relying on it. Nothing published describes jurisdictional coverage of the regulatory corpus or how completeness is assured.

No individual is assessed and two adapted exposures apply. Coverage across more than 100 languages will not be uniform, so market and regulatory intelligence about smaller or less covered jurisdictions rests on thinner and later signal than intelligence about major English language markets, which shapes where institutional attention and capital go.

And social media ingestion as a sentiment source imports the biases of who posts, so measured sentiment reflects the online population rather than the investor base. No coverage breakdown by language or market was located.

No commercial guarantee or indemnity was located, and attribution provides the practical recourse that matters for a research product. Because generated insights are stated to be attributable and fact based, an analyst who doubts a conclusion can retrieve the underlying source and judge it, so an error is discoverable at the point of use rather than after it has entered a decision. That is meaningful in a product whose output feeds investment and compliance judgements. What is absent is anything covering the company's own failure, with no correction or notification process described for a wrong or missed signal.

The underlying model provider is named directly, with the platform stated to leverage leading language models including a specific major commercial model to deliver its generative output, which is a disclosure very few vendors in this index make and which lets a buyer assess that fourth party dependency. Source categories are enumerated clearly across news, social media, regulatory filings and market data, with premium and private sources acknowledged as an augmentation. A major public cloud is identified in third party profiling. What is not disclosed is any individual data supplier or a subprocessor list.

The engine is designed for consumption elsewhere rather than as a destination, with developers named as a user class alongside analysts and agents, and enterprise content integrated into the same search surface so internal and external material are queried together. A named partnership pipes real time market intelligence into a third party's white labelled personalised video briefings, which is the platform being embedded in someone else's client facing product. What is not published is any named financial system integration, and no developer documentation was located.

A major public cloud appears in third party profiling of the technology stack, which is more than nothing, and no region selection, residency commitment or private deployment option is published by the company. Enterprise customers integrating their own internal content into the search index would ordinarily require a stated position on where that content is processed and held.

Two concrete commercial facts are published, which is more than most vendors offer. The founder names the incumbent price point directly, describing continuous intelligence as having been locked behind six figure enterprise contracts and complex integrations, and then states the alternative: a self serve platform where any knowledge worker, developer, enterprise client or agent can access the same engine starting for free. Naming the market's prevailing price and offering a free entry point tells a buyer both what they are escaping and what they can test without procurement. No paid tier pricing is published.

Buyers span financial institutions and corporations, with the newer self serve route extending to individual knowledge workers and developers. Functional coverage is broad across regulatory intelligence, market and macroeconomic insight, conversational enterprise search, chart explanation and, since late 2025, cryptocurrency intelligence. Language coverage above 100 supports genuinely global research. What is not evidenced is the institutional footprint, since no customer, region or segment concentration is disclosed and the company operates from two continents with around eleven people.