Wisesheets has launched a financial data application programming interface and a hosted Model Context Protocol server, expanding the Toronto financial-technology company from spreadsheet-based investment tools into infrastructure for developers, digital financial products and artificial-intelligence agents.

The products, announced July 24, provide programmatic access to company fundamentals that Wisesheets says are collected directly from filings on the U.S. Securities and Exchange Commission’s EDGAR system. The launch places the company in a competitive financial-data market while giving it exposure to rising demand for information services that can be consumed directly by generative-AI applications.

The API distributes standardized income statements, balance sheets, cash-flow statements, financial ratios, historical share prices and company-profile information through a REST interface. Potential applications identified by Wisesheets include stock screeners, discounted-cash-flow models, portfolio dashboards, peer-comparison tools, financial websites and automated research systems.

The accompanying MCP server provides another route into the same data. Model Context Protocol is an open framework through which an AI client can discover and call external tools rather than relying entirely on information embedded in its training data or retrieved from unstructured webpages. After connecting the Wisesheets server, a compatible assistant can request financial statements, compare margins, retrieve prices or screen companies using natural-language instructions.

That distinction is important for financial applications. A general-purpose language model may produce a figure based on stale information, an incorrectly parsed document or an unsupported inference. A tool-connected system can instead obtain a structured value from a dedicated database at the time a user asks the question. The model still interprets the response, but the underlying number comes from an identifiable data service.

Wisesheets is emphasizing filing-level provenance as the main differentiator of its offering. The company says each fundamental field is returned with metadata identifying the original XBRL concept, the relevant SEC accession number and the filing date. Responses can therefore point users back to the disclosure from which a value was extracted.

XBRL, or eXtensible Business Reporting Language, is the structured reporting format used by SEC registrants to label financial facts within regulatory filings. In theory, those tags allow software to locate information such as revenue, operating income or total assets without reading the document in the same way a person would. In practice, companies can use different concepts, extensions and presentation structures, requiring data vendors to map disclosures into consistent categories.

Wisesheets says it uses a proprietary concept-mapping layer to normalize the large number of XBRL tags used by issuers. That normalization is essential for comparative analysis. A screening application cannot reliably rank hundreds of companies by revenue growth, free-cash-flow margin or leverage unless economically similar line items are represented under a stable schema.

The company’s documentation lists coverage of more than 10,400 U.S. stocks and more than 230 financial concepts. It also advertises historical depth of more than 20 years at the upper subscription levels. The service currently focuses on companies filing with the SEC, while expansion into Canada, Europe and the Asia-Pacific region is listed as part of the broader rollout plan.

Wisesheets says new filing information is generally synchronized with its API within approximately 15 minutes. That speed could make the service useful for earnings-monitoring systems that compare newly submitted statements with prior periods, analyst expectations or predetermined risk thresholds. The latency, coverage and validation figures are company claims and have not been independently verified in connection with the launch.

The REST product uses bearer-token authentication, with developers obtaining an API key and passing it through the authorization header. The documentation advises customers to store credentials in environment variables or secrets-management systems rather than hard-coding keys into applications.

The hosted MCP server uses a different connection process. According to Wisesheets, users add the remote server address to a compatible AI application and authorize access through OAuth. That removes the need to paste an API key into the AI client. Documentation provides setup guidance for systems including Claude, ChatGPT, Gemini, Cursor and development environments that support remote MCP connections, although availability can depend on the application and subscription plan.

For fintech developers, the combined approach lowers two separate integration barriers. The REST API offers a familiar interface for production applications, while the MCP service can make the dataset accessible to AI assistants without requiring a developer to create a separate tool wrapper. A team could use the API within a customer-facing dashboard and allow an internal research agent to query the same underlying data through MCP.

The initial pricing is aimed partly at individuals and small-scale builders. Wisesheets offers a free plan with 5,000 API requests per month, up to five years of history and access to annual and quarterly fundamentals, prices, dividends, metrics and ratios. The company says no credit card is required to begin using the free tier, and the MCP server is included.

A $19 monthly Pro plan raises the allowance to 80,000 requests, adds bulk endpoints and provides as much as 15 years of history. A $39 Elite plan includes 200,000 monthly requests and full historical access. The advertised individual subscriptions carry important restrictions: Wisesheets states that they are not intended for company, professional, regulated, fund-management or client-advisory use.

Businesses, registered financial professionals and institutional investment teams are directed to an enterprise plan with custom pricing, commercial-use rights and negotiated request limits. The licensing distinction means the headline consumer pricing is not necessarily representative of the cost for a bank, asset manager, registered adviser or commercial fintech platform.

The launch builds on Wisesheets’ original spreadsheet products, which allow investors to import financial information into Microsoft Excel and Google Sheets. The company says those add-ins have accumulated more than 70,000 users worldwide since 2020. Moving into APIs gives Wisesheets access to a broader customer base because developers can embed data in applications rather than requiring end users to work inside a spreadsheet.

The product also reflects how financial-data distribution is changing. Traditional vendors have generally delivered information through terminals, licensed feeds, downloadable files and proprietary desktop software. Newer providers increasingly offer developer-focused APIs, cloud delivery and usage-based subscriptions. The spread of AI agents is adding another distribution layer in which datasets are exposed as callable tools with descriptions that models can understand.

For investment research, the MCP server could support requests such as retrieving several years of revenue, comparing operating margins across competitors or finding companies that satisfy a group of fundamental conditions. An assistant could then produce a table, narrative or preliminary valuation using the returned values while retaining links to the original filings.

Filing provenance is particularly relevant where research outputs are reviewed by compliance, legal or risk teams. An audit trail can help reviewers reconstruct which disclosure supported a figure and when it was retrieved. It can also make corrections easier when a company files an amended report, changes a taxonomy tag or presents a restatement.

Traceability should not, however, be treated as proof that every analytical conclusion is correct. An AI agent can retrieve an accurate value and still misunderstand whether it represents a quarter, year-to-date period or full fiscal year. It may compare companies with different fiscal calendars, confuse reported and calculated metrics, overlook discontinued operations or combine figures expressed in different units.

Non-GAAP measures present another challenge. Metrics highlighted in earnings releases, including adjusted earnings, organic growth or management-defined free cash flow, may not correspond directly to standardized SEC taxonomy concepts. A fundamentals API built mainly around XBRL statements may require additional processing or separate datasets to reproduce those measures consistently.

Corporate actions and filing revisions also require careful treatment. Stock splits, acquisitions, changes in accounting policy and restated historical periods can alter time-series comparability. Developers building backtests or automated valuation tools will need to understand whether the data is point-in-time, subsequently revised or adjusted using information that was unavailable on the original analysis date.

Wisesheets says its platform includes validation checks intended to detect sign errors, changing units and balance-sheet inconsistencies. The company also describes confidence indicators and filing dates that agents can use when deciding how to present a result. Those controls may reduce errors, but prospective institutional users are likely to test them against primary filings and established datasets before relying on the service in production.

Operational reliability will be another consideration. Applications built on a third-party API depend on its uptime, rate limits, schema stability and response speed. Developers may require service-level agreements, versioning policies, disaster-recovery procedures and advance notice of endpoint changes. Enterprise customers may also seek detailed information on security, logging, data retention and access governance.

The MCP component introduces its own governance questions. Giving an AI assistant access to an external tool can improve accuracy, but organizations must still control which agents can connect, what queries they can make and how retrieved information is used. Tool descriptions, authentication flows and output schemas must be secured against manipulation, while sensitive internal prompts should be separated from unnecessary third-party services.

Wisesheets states that the service supplies data and does not need access to a customer’s proprietary information. Even so, regulated firms evaluating any external AI connector will typically examine authentication, network architecture, vendor controls and whether prompts or query logs are retained. The product’s institutional adoption will therefore depend on more than data coverage and price.

The company is entering a field that already includes established market-data vendors and developer platforms offering fundamentals, prices and company information. Wisesheets’ opportunity is to compete on affordability, ease of integration and the specificity of its source metadata. Its narrower U.S. filing focus could be an advantage for developers prioritizing SEC traceability, but a limitation for products requiring global securities, estimates, fixed income or alternative datasets.

No revenue figures, external funding details or financial forecasts were disclosed with the product announcement. Wisesheets also did not identify initial API or MCP customers. The launch should consequently be viewed as a product expansion rather than evidence that the company has already secured a significant share of the institutional financial-data market.

Still, the release illustrates a broader fintech trend: financial information is being packaged not only for analysts and applications but also for autonomous or semi-autonomous software agents. As those systems become more involved in research, portfolio monitoring and client-service workflows, data vendors are likely to compete increasingly on machine-readable provenance, permissioning and auditability.

Wisesheets’ API and MCP server are designed around that shift. Their commercial impact will depend on whether the company can deliver consistent normalization and dependable infrastructure at scale. For developers, the immediate appeal is simpler access to structured SEC fundamentals. For financial institutions, the more consequential question is whether the service can satisfy the accuracy, licensing and governance standards required when AI-generated analysis informs real investment decisions.