Programmable Infrastructure for the Agent Era: VAST Data AI OS API Documentation Hub
One front door for Cluster and DataEngine APIs — with search, machine-readable indexes, and one-click LLM handoffs on every endpoint.
VAST AI OS 5.5 unified hyperscale vector retrieval, native SQL analytics, managed pipeline compute, and zero-copy block operations on one platform. The VAST AI OS API layer underneath ties it all together — but finding the right endpoint used to mean hunting across separate guides, in-cluster Swagger pages, and release notes.
That changes today. Clicking API Documentation on kb.vastdata.com now opens a dedicated space with overview and getting-started guides, clearly documented VAST Cluster 5.5 endpoints, and improved navigation and search.
One front door for every API surface
The new hub unifies two API families that power VAST Data AI OS:
VAST Cluster API — programmatic access to the VAST Management Server (VMS), the same administrative layer behind the Web UI and CLI. Query cluster health, provision views and snapshots, manage tenants and quotas, pull performance metrics, and automate day-two operations over standard HTTPS with JSON responses.
VAST DataEngine API — the compute orchestration surface for pipelines, functions, and triggers. Deploy DataEngine on a tenant, wire up event brokers and Kubernetes clusters, and manage telemetry collection and query services for logs and traces across your pipeline fleet.

Figure 1. The VAST Knowledge Base homepage now links directly to the unified API Documentation space.

Figure 2. The API landing page introduces the VAST Cluster 5.5 API with overview content and a structured endpoint catalog.
Whether you are scripting backups, building GitOps controllers, or wiring AI agents into your storage fabric, you start from the same home.
Built for discovery — by humans and agents
Modern infrastructure teams do not only write scripts. They deploy AI coding assistants, autonomous agents, and MCP servers that need to find, understand, and call APIs without a human clicking through menus.
The new documentation space is designed for that workflow:
- Structured navigation groups hundreds of v5.5 endpoints under logical categories — Infrastructure, Element Store, Data Protection, Database, Data Engine, and more.
- Global search (Ctrl+K) surfaces API endpoints, procedural articles, and categories from a single query.
- Machine-readable index — the site publishes a complete documentation index at
https://kb.vastdata.com/llms.txt, giving AI assistants a canonical map of every page. - Per-page markdown — every API article is also served at a
.mdURL for programmatic fetch and RAG indexing.
One-click LLM handoffs on every endpoint
Each API reference page includes an Article options menu (⋯) with four LLM-ready actions:
- Copy page — copies the full endpoint as Markdown for LLMs: description, method, path, auth, OpenAPI schema, parameters, and response models.
- View as Markdown — opens plain-text markdown in a new tab (append
.mdto the page URL). Fetch withcurl, mirror in CI, or index for RAG. - Open in ChatGPT — starts a ChatGPT session with the page pre-loaded as context for sample code, pagination, or error handling questions.
- Open in Claude — same one-click handoff to Claude, grounded in the endpoint you are viewing.

Figure 3. The Getting Started guide covers authentication, API versioning, pagination, and request formation.

Figure 4. Each endpoint is documented with HTTP method, path, parameters, request/response schemas, and examples.

Figure 5. The DataEngine API section covers provisioning, telemetry collection, and telemetry query services.

Figure 6. Global search returns API endpoint references alongside related articles.

Figure 7. The Article options menu — copy as LLM-ready markdown, view raw markdown, or hand the page to ChatGPT or Claude.

Figure 8. One click on Open in ChatGPT starts a session with the endpoint markdown pre-loaded — here ChatGPT explains the List View Policies API from the docs.
This pairs with VAST Admin MCP — an open-source MCP server for natural-language cluster management in Cursor, Claude Desktop, and VS Code. Read-only by default; write access is explicit. The API hub is the reference both tools and operators rely on. See also our earlier walkthrough: Talk to Your Data Infrastructure with MCPs.
What you can automate — and why it matters now
VAST AI OS 5.5 is not a storage array with an API bolted on. It is a data platform where storage, database, and compute share one namespace.
- AI and RAG pipelines. DataEngine functions ingest, transform, and index data on-platform. The Engine API provisions triggers, deploys functions, and queries pipeline telemetry.
- Agent sandboxes and Kubernetes. Managed Kubernetes on CNodes plus API-addressable block volumes, instant clones, and snapshots enable zero-copy workspace fan-out and durable agent state.
- Analytics and governance. The Cluster API exposes DataBase settings, catalog queries, and tenant metrics. API tokens with RBAC, tenant isolation, and enforced pagination (v8) give platform teams guardrails for automation.
Getting started in minutes
- Form requests as
https://<VMS_VIP>/api/v8/<object>/for 5.5 endpoints. - Authenticate with basic auth or API tokens over HTTPS.
- Paginate list endpoints with
pageandpage_size— enforced at scale in 5.5. - Use the endpoint catalog,
vastpy, or OpenAPI-generated clients.
For AI workflows: index llms.txt, copy endpoint markdown into your agent, or use VAST Admin MCP with scoped tokens.
The platform is the API
Your data platform should be as programmable as the applications running on it. With AI OS 5.5 generally available and this documentation hub live, that programmability is easy to find, search, and hand to CI pipelines, internal portals, or autonomous agents.
Explore the new API documentation at kb.vastdata.com/api-5-5/apidocs.
Related on Lots of Data
Related on Lots of Data: VAST AI OS 5.5 Is Generally Available · Event Streaming Is the Backbone of Real-Time AI · What Does VAST Data Actually Do? · VAST CSI on OpenShift
