chroma-cloud

Instructions

Intake

Do not block on a long questionnaire. Ask only for details that are missing and required to choose the right path:

• Dense only or hybrid search
• Whether CHROMA_API_KEY, CHROMA_TENANT, and CHROMA_DATABASE are already configured
• Existing embedding choice, if any

If the user has no embedding preference, default to Chroma Cloud Qwen. If hybrid search is required, use Schema() and Search(). If the task is narrow, such as fixing an existing query, reviewing code, or answering an API question, proceed with the repo context instead of forcing intake.

What to validate

• Correct client import (CloudClient vs Client)
• Environment variables are set for Cloud deployments
• Embedding function package is installed when the selected TypeScript embedding requires one
• Schema() and Search() are only used for Cloud workflows
• Important: get_or_create_collection() accepts either an embedding_function OR a schema, but not both. Use schema when you need multiple indexes, hybrid search, or sparse embeddings; use embedding_function for simple dense-only search.

Quick Start

Use the CLI topic to authenticate and write Cloud credentials:

chroma login
chroma db create <db_name>
chroma db connect <db_name> --env-file

Then create a CloudClient and choose the API based on the search mode:

import { CloudClient } from 'chromadb';

const client = new CloudClient();
const collection = await client.getOrCreateCollection({ name: 'my_collection' });

Use collection.query() for dense-only search. Use Schema() plus Search() only when the user needs hybrid retrieval, multiple indexes, or more expressive ranking/query composition.

Cloud Guidance

Collections are the main isolation boundary in Chroma Cloud, and metadata is the main filtering mechanism inside a collection. Reach for Schema() only when you need explicit dense+sparse or multi-index configuration, and reach for Search() only when query() is not expressive enough.