Developer Infrastructure · Dino

Context, intelligence and retrieval

A continuously evolving knowledge base of each API's structure, behaviour and history, served to developers and agents by a retrieval engine built on knowledge graphs, semantic search and contextual ranking.

Founder · Architect · Full-Stack Engineer

The problem

An agent handed a raw OpenAPI file or a whole repository has too much and too little at once: too much text, and too little knowledge of how the API actually behaves. Good agent work starts with the right context, in the right amount.

The knowledge base

The Dino Context Graph is a typed, versioned model of an API: its entities, operations, flows and behaviours, from authentication and error conventions to versioning. Every fact carries where it came from: observed, declared or inferred. Each scan freezes the graph and its findings into an immutable snapshot. Per-operation memory tracks latency and error baselines, detects schema drift, and keeps degraded runs out of the baselines.

The retrieval engine

Dino serves that knowledge through deterministic, scoped retrieval:

a knowledge graph and a full-text index working together;
hybrid retrieval with fusion and contextual reranking;
bounded context slices, each with a manifest of what was selected and why.

The same question gets the same context, sized for the task, every time.

System

API discovery

Dino finds every operation an API exposes, REST and GraphQL, with its schemas, authentication requirements, dependencies and relationships, and keeps that inventory current.

DiscoveryOpenAPIGraphQL

System

API understanding

Dino interprets API structures, identifies relationships, reconstructs workflows, connects technical operations to their business meaning, and writes the documentation teams wish they had.

UnderstandingDocumentation

System

Built into the agent: MCP, CLI, skills and assistant

Dino puts API knowledge, context retrieval, findings and quality checks directly inside coding agents through MCP, with a developer CLI, agent skills and a natural-language API assistant.

MCPCLIAI agents

Context, intelligence and retrieval

The problem

The knowledge base

The retrieval engine

Related systems

API discovery

API understanding

Built into the agent: MCP, CLI, skills and assistant

The problem

The knowledge base

The retrieval engine