Chroma

Embedded open-source vector database using HNSW with squared L2 distance on a single node.

Provider: default
Kind: Vector database
Category: Data
Official docs: https://docs.trychroma.com/docs/collections/configure

Properties · 6 set by this component

property	type	values	required	kind default	this component
deployment · component property	enum	single-node \| distributed \| cloud	no	single-node	inherited
dimensions	integer	—	no	—	inherited
distanceMetric	enum	cosine \| euclidean \| dot-product	no	—	▸ euclidean
efConstruction · component property	integer	—	no	100	inherited
hybridSearch	boolean	—	no	false	▸ true
indexType	enum	flat \| hnsw \| ivf \| disk-ann \| lsh	no	—	▸ hnsw
managed	boolean	—	no	false	▸ false
maxNeighbors · component property	integer	—	no	16	inherited
metadataFiltering	boolean	—	no	false	▸ true
namespaces	boolean	—	no	false	inherited
quantization	enum	none \| scalar \| product \| binary	no	none	▸ none
replicas	integer	—	no	—	inherited
space · component property	enum	l2 \| cosine \| ip	no	l2	inherited

Rules that apply · 0

Every rule whose selector covers this component's kind or its category. Whether a rule fires depends on the design: its properties, its edges and its groups.

No rule reads this kind yet. That is a gap, not a guarantee.

Use it in a design

nodes:
  - { id: "chroma", component: "chroma" }

Other Vector database products · 23

Alibaba Cloud Milvusalibaba-milvusalibabacloudvector-database Amazon S3 Vectorsaws-s3-vectorsawsvector-database Agent Memorycloudflare-agent-memorycloudflarevector-database Cloudflare Vectorizecloudflare-vectorizecloudflarevector-database Deep Lakedeeplakevector-database Epsillaepsillavector-database KDB.AIkdb-aivector-database LanceDBlancedbvector-database Marqomarqovector-database Milvusmilvusvector-database MyScalemyscalevector-database pgvectorpgvectorvector-database pgvectorscalepgvectorscalevector-database Pineconepineconevector-database Qdrantqdrantvector-database turbopufferturbopuffervector-database Valdvaldvector-database Generic vector databasevector-databasevector-database Vespa Cloudvespa-cloudvector-database Weaviateweaviatevector-database Zilliz Cloudzilliz-cloudvector-database DigitalOcean Vector Databasesdigitalocean-vector-databasedigitaloceanvector-database Vertex AI Vector Searchgoogle-vertex-ai-vector-searchgcpvector-database

Source

1# yaml-language-server: $schema=../../schema/component.schema.json
2# Chroma: Apache-2.0 embedded vector database. "Configuring Chroma collections": the default space is
3# squared L2, single-node uses an HNSW index with ef_construction 100 and max_neighbors 16 while the
4# distributed and Cloud deployments use SPANN. "Introduction": metadata filtering and full-text search.
5schemaVersion: "1"
6type: "component"
7id: "chroma"
8kind: "vector-database"
9provider: "default"
10docs: "https://docs.trychroma.com/docs/collections/configure"
11defaults:
12  indexType: "hnsw"
13  distanceMetric: "euclidean"
14  quantization: "none"
15  metadataFiltering: true
16  hybridSearch: true
17  managed: false
18properties:
19  deployment:
20    type: "enum"
21    values: ["single-node", "distributed", "cloud"]
22    default: "single-node"
23  space:
24    type: "enum"
25    values: ["l2", "cosine", "ip"]
26    default: "l2"
27  efConstruction:
28    type: "integer"
29    default: 100
30  maxNeighbors:
31    type: "integer"
32    default: 16

The same page as markdown, for an agent: /catalogue/components/chroma.md