OSS-first docs
These docs teach the open system first: contracts, generated surfaces, runtimes, governance, and incremental adoption. Studio shows up as the operating layer on top, not as the source of truth.
Qdrant integration for vector search and embeddings storage.
Source Definition
export const qdrantIntegrationSpec = defineIntegration({
meta: {
key: 'vectordb.qdrant',
version: '1.0.0',
category: 'vector-db',
title: 'Qdrant Vector Database',
description: 'Qdrant integration for vector search and embeddings storage.',
domain: 'ai',
owners: ['platform.ai'],
tags: ['vector-db', 'search'],
stability: StabilityEnum.Experimental,
},
supportedModes: ['managed', 'byok'],
capabilities: {
provides: [
{ key: 'vector-db.search', version: '1.0.0' },
{ key: 'vector-db.storage', version: '1.0.0' },
],
requires: [
{
key: 'ai.embeddings',
optional: true,
reason:
'Required if vectors are generated via hosted embedding services',
},
],
},
configSchema: {
schema: {
type: 'object',
properties: {
apiUrl: {
type: 'string',
description:
'Base URL for the Qdrant instance (e.g., https://qdrant.example.com).',
},
collectionPrefix: {
type: 'string',
description:
'Prefix applied to all collection names for this tenant.',
},
},
},
example: {
apiUrl: 'https://qdrant.example.com',
collectionPrefix: 'tenant_',
},
},
secretSchema: {
schema: {
type: 'object',
properties: {
apiKey: {
type: 'string',
description: 'API key or token when authentication is enabled.',
},
},
},
example: {
apiKey: 'qdrant-api-key',
},
},
healthCheck: {
method: 'ping',
timeoutMs: 4000,
},
docsUrl: 'https://qdrant.tech/documentation/quick-start/',
constraints: {
quotas: {
collections: 100,
pointsPerCollection: 1_000_000,
},
},
byokSetup: {
setupInstructions:
'Provide the HTTPS endpoint of your Qdrant cluster and generate an API key with read/write access to the collections that will be managed.',
},
});