Whatever message this page gives is out now! Go check it out!
Chroma vector store configuration options
Last update:
May 18, 2026
This document describes settings when Chroma is selected as the vector store provider (where embeddings and metadata are stored and queried). Chroma does not produce embeddings: configure the embedding model separately (for example under AI Services > Embedding models, or via Source / Provider in the Embedding model section).
Cross-check field names and defaults against your product build and Chroma documentation.
Global behavior notes
- Embedding dimension: Chroma collections are created with an embedding function or vector dimension implied by your pipeline. Ensure the embedding model you attach outputs a dimension consistent with how this integration creates or updates the collection—mismatches cause runtime errors when adding or querying vectors.
- Secrets: The API key (when used) authenticates to Chroma Cloud or a secured deployment. Store saved configurations like any other secret.
- Local vs cloud: Chroma Cloud typically requires an API key. A local Chroma server without auth may leave the API key empty.
- Hierarchy: Tenant → database → collection provides increasing isolation. Use non-default names when you run multi-tenant or multi-database setups; otherwise default is fine for a single application.
Basic configuration
Field | Description |
|---|---|
Provider | Chroma. The selected vector store provider for this configuration. |
URL | Required. Base URL for the Chroma HTTP API. For a local server, often http://localhost:8000. For cloud or remote hosts, use https://<host>:<port> (or the URL your operator provides). |
API key | Optional for local instances without authentication. Required for typical Chroma Cloud (or secured) deployments. Create and rotate keys in your Chroma Cloud project; treat stored keys as secrets. |
Collection name | Optional with a product default (often default). Logical collection for vectors. If it does not exist, the product may create it automatically—confirm behavior in your release notes. |
Database name | Optional with a default (often default). Separates data within a tenant when your Chroma deployment supports multiple databases. |
Tenant name | Optional with a default (often default). Top-level isolation for multi-tenant Chroma (tenant contains databases; databases contain collections). Change only when your deployment uses explicit tenants. |
Call timeout (ms) | Maximum time to wait for a Chroma API call. Default is often 60000 ms. A practical range is often 10000–300000 ms depending on batch size and network latency. |
Top K | Maximum number of results to return from a similarity search. Must be a positive integer when set. |
Min score | Optional minimum similarity score; results below the threshold are dropped. Often interpreted between 0 and 1; leave empty to return all top-K matches without a score cutoff. |
---
Embedding model (related UI)
Field | Description |
|---|---|
Source | None — skip embedding configuration only when valid for this flow. Pre-configured — use a profile from AI Services > Embedding models. Custom — configure embedding inline. |
Provider | After you choose a source, select the embedding provider (OpenAI, Mistral, Ollama, and so on). That drives vector generation; Chroma settings define where vectors are stored and how search is scoped (collection, database, tenant). |
---
We're glad. Tell us how this page helped.
We're sorry. Can you tell us what didn't work for you?
Share additional feedback to help us improve.
0/255
|
