Book a Demo Talk to Sales

Developers · One runtime

Build enterprise AI apps on one runtime

Typed graph, hybrid retrieval and permission-aware AI in a single deployable. No gluing four services together and owning the seams for the next three years.

Start Free Read the Docs

Languages C# SDKs and REST APIs Runtime One deployable Permissions Resolved at query time Models Any model, hosted or local

One runtime · your servers

Graph, retrieval and models in one runtime Three stacked layers inside a single frame: the language model of your choice on top, hybrid retrieval beneath it, and a graph of connected records underneath both, with one permission boundary around all three. ONE DEPLOYABLE · ONE PERMISSION BOUNDARY Models you chooseOpenAIAnthropicMetaLocal Hybrid retrievalKeywordVectorGraph walk Context graph
Models, retrieval and the graph in one runtime Three stacked layers inside one frame: the models you choose, hybrid retrieval beneath them, and the context graph they both stand on. ONE DEPLOYABLE Models you chooseHostedLocalHybrid retrievalKeywordVectorContext graphRecordsLinks

Install it. Build on it.

  1. 01 Install it free
  2. 02 Connect a system
  3. 03 Build on the graph

One deployable. One permission model.

Install · Free

Run Studio on your own machine

One command. On macOS and Linux Studio runs in a container, on Windows through its installer. It runs on the machine you install it on, and so does your data.

macOS · Linux

01 · Install

Starts Studio as a container named curiosity, with Docker or Podman, whichever you have. Without either it tells you how to install one, and changes nothing. Run it again to update: your data stays in its volume.

Terminal
 curl -fsSL https://curiosity.ai/install.sh | bash

02 · Day to day

Open localhost:8080 and sign in as admin with the password the script prints. The container is yours to manage. With Podman, swap docker for podman.

Terminal
 docker logs -f curiosity
 docker stop curiosity
 docker start curiosity
 docker restart curiosity

Read the script · Running on Docker

Windows

01 · Install

Downloads the Windows installer, checks its signature and runs it for your account. Start Studio from the Start Menu, open localhost:8080 and sign in as admin / admin. Change that password first.

PowerShell
 irm https://curiosity.ai/install.ps1 | iex

02 · As a Windows service

The script sets up no service. When you want one, open an elevated prompt, install to a folder outside C:\Users and run the guided setup there. It asks for the data folder, the admin account and the account the service runs as.

PowerShell
 $env:CURIOSITY_INSTALL_DIR = "E:\Curiosity"
 irm https://curiosity.ai/install.ps1 | iex
 cd E:\Curiosity
 .\configure-windows-service.ps1 setup

Read the script · Windows installation

Trusted in complex environments

Build on one framework

Everything runs together, so there is nothing to integrate

The pieces ship as one process. That is the whole argument.

01

One deployable

Graph, retrieval, enrichment and model routing in a single runtime you can run on a laptop or a cluster.

02

One permission model

Access is resolved once, at query time, and every surface inherits it. There is no second place to get it wrong.

03

One set of types

The schema the graph enforces is the schema your code sees and the schema the model is handed.

04

Your language

C# SDKs and REST APIs. Version it, review it, test it like the rest of your code.

Capabilities

What the runtime gives you

Nine surfaces, all against the same graph and the same permissions.

Knowledge graph

Platform

A typed, in memory graph built for traversal at interactive speed across tens of terabytes.

Learn More

Data enrichment

Platform

Extract entities, classify records and normalize identifiers as data lands, not in a nightly batch.

Learn More

Integrations

Platform

70+ connectors, and an interface for the system nobody else has heard of.

Learn More

Search

Retrieval and agents

Keyword, vector and graph traversal in one query, with permissions applied at query time.

Learn More

Embeddings

Retrieval and agents

Generate and store vectors next to the graph, so retrieval does not need a second datastore.

Learn More

AI agents

Retrieval and agents

Agents that get typed context and an audit trail rather than a scraped text dump.

Learn More

Extensibility

Retrieval and agents

Custom types, custom pipelines and your own logic where the defaults do not fit.

Learn More

Developer tools

Build and run

C# SDKs, REST APIs and a build IDE, so this is code you can review and version.

Learn More

Security

Build and run

Relationship-based permissions, SSO with enforced MFA and a full audit trail.

Learn More

In production

Numbers from live deployments

Customer Stories

30TB+

Connected across enterprise systems in production

20K+

Active users across enterprise deployments

70+

Enterprise systems with a connector

Integrations

70+ enterprise systems, connected

Tickets, part data, SharePoint, Jira, Confluence and S3, with the permissions each source keeps. An interface covers the system nobody else has heard of.

See Integrations

Models

Your model, swapped as configuration

OpenAI, Anthropic, Mistral, Meta or local weights. Curiosity is not the model. It is what gives the model context.

See Models

Security

Built for security

Made in Germany and built for data sovereignty. Your data stays in your environment, encrypted and access controlled down to the individual record.

Visit Our Trust Center Book a Demo

On premises or private cloud

Deploy on your own infrastructure or in certified data centers in Europe.

Single sign on and enforced MFA

Access synced to your data sources with relationship based permissions.

Full audit trail

Every query, download and data access logged.

Data residency

Your data stays in the region you choose.

Made in Germany

Built in Munich.

GDPR compliant

European data protection standards.

Your data. Your infrastructure.

Build it on your own data

Book a Demo Talk to Sales