Build enterprise AI apps on Curiosity Workspace

The graph + search + AI platform for building AI applications on your own data.

Build your first app (30-minute tutorial)Read the architecture overview

Curiosity Workspace for developers

This section is for developers integrating Curiosity Workspace into a product. For installing, securing, and operating the platform, see Deploy and manage. For building custom interfaces, endpoints, and AI tools on top of an existing workspace, see Workspace Customization.

Curiosity Workspace is a single, schema-driven platform that combines three layers — graph, search, and AI — into a unified runtime. Permissions (ReBAC) and relevance are enforced consistently across all three.

You build applications on Workspace by:

Defining a schema for the entities and relationships in your domain.
Writing connectors to keep that data fresh.
Configuring search (text, vector, hybrid) and embeddings for the fields you want findable.
Exposing custom endpoints and AI tools that wrap retrieval, grounding, and business logic.
Building a custom interface — or using the built-in one.

The 30-minute developer journey

A complete end-to-end walkthrough that takes you from docker run to a permission-aware AI app:

Build your first enterprise AI app

If you're new, also read:

Developer prerequisites — what to install and configure before you start.
Concepts for developers — a 5-minute orientation to the vocabulary used in these docs.
Quickstart — the shorter version: get Workspace running locally.

The platform

Graph

Typed nodes and edges with stable keys and explicit schemas. Edges are first-class for navigation, faceting, and constrained retrieval.

Search

Text + vector + hybrid retrieval, with property facets and graph-relationship facets. Every query is filtered by user permissions before it returns.

AI

Embeddings, entity extraction, LLM orchestration, AI tools, and agents. Grounded in graph + search retrieval, never untethered.

What you'll build with

Data Connectors

C# or Python connectors using Curiosity.Library. Stable-key upserts, graph edges, incremental sync, ACL ingestion.

Custom Endpoints

Server-side C# endpoints with full access to the graph, search, and AI runtime. Token-scoped and permission-aware.

AI Tools

Annotated C# classes the LLM can invoke during chat. Permission-aware tool calls with citations and snippets.

Custom Interfaces

Tesserae-based front-ends that run as part of the Workspace. Reuse built-in chat, search, and graph components.

Curiosity Workspace Home View

Read by topic

Fundamentals

Architecture

The unified graph + search + AI runtime: components, request flows, deployment model, scaling.

Data Flow

From raw source record to graph node, search index, and grounded AI answer.

Access Control

The ReBAC model: graph-path permissions, ACL ingestion, and query-time enforcement.

Development & integration

Connectors

Methods for ingesting data from external sources.

API Overview

Interacting with workspace services programmatically.

Custom Endpoints

Extending the platform with custom server-side logic.

Examples and patterns

Examples gallery

Production-shaped blueprints for technical support, manufacturing, compliance, and customer 360. Plus migration guides from Elasticsearch+LangChain and from legacy enterprise search.

Tutorials

Walk-throughs of the technical-support sample and the HackerNews demo, plus domain-specific recipes.

Operations, security, and customization

Deploy and manage

Install, secure, monitor, back up, and upgrade Curiosity Workspace.

Customize

Custom interfaces, endpoints, AI tools, NLP pipelines, and connectors.

Need Help?

Community Support: Ask questions and share your projects.
Report Issues: Found a bug? Let us know.