Monospace from Directus

Product Introduction

1. Definition: Monospace from Directus is a governed API layer, a middleware platform that acts as a real-time, read-write data fabric. It connects disparate data sources—including SQL/NoSQL databases, REST APIs, and SaaS applications—into a single, unified GraphQL/REST API endpoint without moving or copying the underlying data.
2. Core Value Proposition: It exists to solve the critical tension between developer velocity and enterprise data governance. Monospace provides instant, governed data access for developers, business teams, and AI agents, enabling rapid application and AI workflow development on top of legacy and modern systems while maintaining centralized security, permissions, and audit controls.

Main Features

1. Universal Data Connectors with Real-Time Introspection: Monospace features a library of connectors that establish live links to source systems. Unlike ETL tools, it performs real-time schema introspection and query translation, dynamically understanding table structures, relationships, and API endpoints. This allows it to generate accurate data models and APIs on-the-fly without pre-defined schemas or data duplication.
2. Granular, Query-Level Governance Engine: The platform's core differentiator is its ability to enforce permissions within the data query itself. It goes beyond gateway-level allow/deny by injecting row-level security (RLS) and field-level masking directives directly into queries sent to the source database or API. This ensures data filtering happens at the source, and every user, app, or agent only sees authorized data.
3. Unified Agentic Access via MCP (Model Context Protocol): Monospace provides a single, governed bridge for AI agents to enterprise data. It exposes all connected data sources through a standardized MCP server, allowing AI assistants (like those in Claude or Cursor) to perform read/write operations. This is coupled with "Skills"—configurable prompts and constraints that teach agents the specific data schema and safe interaction patterns, keeping actions within guardrails.

Problems Solved

1. Pain Point: The "data access bottleneck." Developers and data teams waste weeks waiting for database access, custom API endpoints, or sanitized data dumps, severely slowing innovation cycles.
2. Target Audience: Enterprise development teams, platform engineers, CIOs/CTOs managing data governance, and AI/ML engineers building agentic workflows. Specific personas include the Full-Stack Developer needing live product data, the Business Analyst building internal tools, and the AI Engineer creating data-aware agents.
3. Use Cases: 1) Rapid internal tool development (e.g., a customer support dashboard pulling from PostgreSQL, Salesforce, and Zendesk). 2) Safely enabling AI agents to execute business processes (e.g., an agent checking inventory and creating a shipment). 3) Modernizing legacy applications by exposing their database through a secure, modern API without rewriting the backend.

Unique Advantages

1. Differentiation: Unlike data warehouses (Snowflake, BigQuery) that copy data for analytics, Monospace queries live operational data. Unlike generic API gateways or MCP gateways, it enforces granular permissions within queries, not just at the connection point. Unlike iPaaS (Zapier, Workato) that rely on async pipelines, it enables synchronous, request-time data access.
2. Key Innovation: The "Governed API Layer" architecture itself. The technical innovation is the real-time fusion of a dynamic API generator with a deeply integrated, query-level policy engine. This allows it to present a unified, immediately usable interface from any data source while maintaining control that is typically only possible with a single, homogenous database.

Frequently Asked Questions (FAQ)

1. How does Monospace handle data security and permissions compared to a traditional API gateway? Monospace enforces security at the data layer, not just the network layer. While an API gateway can authenticate and route requests, Monospace injects row-level and field-level security constraints directly into the SQL or API call sent to the source system, ensuring the principle of least privilege is enforced by the data source itself.
2. Can Monospace connect to legacy or on-premises databases? Yes, Monospace is designed for hybrid and multi-cloud environments. It can connect to on-premises databases using secure methods like SSH tunneling or via a lightweight connector agent installed within the private network, ensuring no data ever needs to be moved to the public cloud for access.
3. What is the performance impact of adding Monospace as a layer between my app and database? Monospace is designed as a thin orchestration and governance layer. The primary performance impact is the network hop to the Monospace instance. The query execution itself is performed by the source database, leveraging its existing indexes and optimizations. For complex queries joining data across sources, Monospace performs in-memory federation, which requires careful planning for large datasets.
4. How does Monospace for AI agents differ from giving an AI direct database access? Monospace provides a critical abstraction and safety layer. Instead of giving an AI agent raw database credentials, it connects via a controlled MCP server with a dedicated identity. The agent's actions are limited by predefined "Skills" and are subject to the same granular permissions (row/field level) and comprehensive audit logging as human users, preventing unrestricted, high-risk queries.