Financial institutions are building armies of autonomous agents to execute operational workflows. As the ratio of AI agents to human employees inverts, traditional compliance QA breaks. You cannot manually sample 10% of alerts when an AI is processing millions of transactions, disputes, and marketing assets.

The bottleneck isn’t building the AI; it is proving to your Chief Compliance Officer that your AI is safe to deploy.

Today, we are officially launching the solution to that bottleneck: The Midlyr API Platform.

The Midlyr API is the compliance primitive for the agentic enterprise. We are providing internal engineering teams with the exact regulatory data and governance controls they need to safely build, screen, and deploy AI agents in production.

The platform is powered by two core engines:

1. The Regulation API (The Data Engine)

The hardest part of building compliance AI isn’t the foundational model; it is the data engineering. Financial regulations are a messy dependency graph spanning seven layers of text (Statutes, Regulations, SROs, Exam Handbooks, Agency Guidance, etc.).

If your engineering team is building web scrapers to parse PDFs from the FDIC or the CFPB, you are wasting cycles on undifferentiated heavy lifting.

The Regulation API solves this. We provide a continuously updated, highly structured, and fully indexable library of banking law. Whether you are building an internal RAG application or a regulatory discovery agent, our API allows you to query, filter (by jurisdiction or authority), and retrieve the exact plain-text excerpts you need, instantly.

2. The Screen Analysis API (The Governance Engine)

If the Regulation API is the data, the Screen Analysis API is the guardrail.

When your internal agent drafts a piece of text—whether it is a response to an electronic fund transfer dispute, a debt collection dunning letter, or a new marketing asset—how do you mathematically prove it is compliant?

Instead of relying on generalized LLM “vibes,” developers pass the agent’s payload and a scenario tag to our Screen Analysis API. Our engine orchestrates a multi-step analysis against our proprietary regulatory graph. It returns a structured JSON output featuring a quantitative Risk Score, prioritized findings (e.g., P1 Blocking / P2 Material), and the exact, traceable regulatory citations to justify the decision.

Three Ways to Integrate

We believe compliance infrastructure should adapt to your stack, not the other way around. We are launching with three distinct integration paths to meet developers exactly where they are:

  • The TypeScript SDK (@midlyr/sdk-js) [Preview]: For teams building Node or TypeScript applications, our official SDK provides full type-safety for all payloads and regulatory schemas, giving you direct access to the REST endpoints. (Python and Go SDKs coming soon). Explore the SDK Docs →
  • The CLI (@midlyr/cli): For teams that want to execute compliance checks directly from their terminal or integrate Midlyr into their existing CI/CD pipelines without writing custom integration code. Explore the CLI Docs →
  • The MCP Server (Model Context Protocol): For teams already building with advanced agentic frameworks or using IDEs like Cursor and Claude Code. By connecting the Midlyr MCP server, your existing agents instantly inherit native tools to browse the regulatory library or trigger compliance screenings. Explore the MCP Docs →

Move Fast and Stay Compliant

You no longer have to choose between deploying AI rapidly and protecting the institution from regulatory risk.

The Midlyr API gives your developers the regulatory data they need to build, and the screening governance they need to deploy. The foundational models are ready. Now, the infrastructure is too.

Ready to start building? Get started for free on the Midlyr Developer Portal →