MCP server
Chroma MCP server
Chroma's official MCP server lets AI create collections, add, query, update and delete documents in a Chroma vector database, local or cloud.
- Official server
- Databases
At a glance
- Maintained by
- Chroma
- Tools exposed
- 10 listed below
- Reviewed
- September 24, 2026
Why it matters
What it does and why it matters
Chroma is an open-source vector database that is easy to run locally and also available as a cloud service. Its official MCP server gives an assistant full control of collections and documents: create, add, query, update and delete. It works with a temporary in-memory database, a local folder or Chroma Cloud.
That makes it a practical choice for prototypes and small knowledge assistants, where a team wants search by meaning without running heavy infrastructure.
Because it has full read and write access, including deleting whole collections, we keep it to local or development data during building. For production use, a curated collection with a separate loading process and search-only access through your app is safer.
Use cases
Best business use cases
Local prototypes
Assistant memory
Document search
RAG testing
Cloud collections
Tools
Available tools
Tool names as published by the maintainer at review time. Versions change, so check the repository for the current list.
Read and write: Full read and write, including deleting documents and collections.
-
chroma_list_collectionsLists collections with paging.
-
chroma_create_collectionCreates a collection with optional HNSW settings.
-
chroma_peek_collectionShows sample documents from a collection.
-
chroma_get_collection_infoGets collection details.
-
chroma_add_documentsAdds documents with optional metadata and IDs.
-
chroma_query_documentsRuns semantic search with filters.
-
chroma_get_documentsGets documents by ID or filter.
-
chroma_update_documentsUpdates document content, metadata or embeddings.
-
chroma_delete_documentsDeletes specific documents.
-
chroma_delete_collectionDeletes a collection.
Workflows
Example workflows
Test a retrieval idea locally
-
1
Start the server with a local persistent folder.
-
2
Create a collection for sample documents.
-
3
Add documents through the assistant.
-
4
Run real questions as queries.
-
5
Adjust chunking and repeat.
Clean up a collection
-
1
Peek at the collection contents.
-
2
Find outdated documents by metadata.
-
3
Review the list the assistant returns.
-
4
Approve updates or deletions.
-
5
Re-run test queries.
Prompts
Example prompts
“Create a collection called policies and add these three documents.”
“Query the policies collection for "remote work equipment" and show the top five.”
“List documents in the collection updated before 2025.”
Setup
Installation and configuration
Examples for Claude Desktop and Claude Code. Replace the placeholder values with your own keys, and store secrets outside version control.
Claude Desktop: claude_desktop_config.json
{
"mcpServers": {
"chroma": {
"command": "uvx",
"args": [
"chroma-mcp",
"--client-type",
"cloud",
"--tenant",
"your-tenant-id",
"--database",
"your-database-name",
"--api-key",
"your-api-key"
]
}
}
}Claude Code: terminal
claude mcp add chroma -- uvx chroma-mcp --client-type persistent --data-dir /full/path/to/your/data/directory- Local server: stdio. PyPI: chroma-mcp (uvx chroma-mcp). Client types: ephemeral, persistent, http, cloud.
- Needs: None for ephemeral or persistent local use; Chroma Cloud: tenant, database, API key (CHROMA_TENANT, CHROMA_DATABASE, CHROMA_API_KEY); Optional embedding provider keys: CHROMA_<PROVIDER>_API_KEY.
Security
Security considerations
-
Development data
Use local or development collections while building.
-
Review deletes
Confirm before deleting documents or collections.
-
Cloud credentials
Scope Chroma Cloud keys to one database.
-
Personal data
Avoid storing personal data in test collections.
Related servers
Servers that pair well with it
Qdrant
Qdrant's official MCP server gives AI a semantic memory: store information in a Qdrant collection and find it again by meaning.
Pinecone
Pinecone's official developer MCP server lets AI search Pinecone docs, manage indexes, upsert records, search and rerank.
Weaviate
An MCP server built into Weaviate that lets AI read collection settings, list tenants, run hybrid search and, when allowed, upsert objects.
Keep exploring
Solutions and services that use it
Related services
View all related services- MCP integration Connect AI assistants to your CRM, files, databases and tools through Model Context Protocol servers, with safe permissions.
- Database design Data models that stay fast and correct as your business grows, with backups and access rules in place.
- AI Product Development AI agents, knowledge assistants, copilots and document automation built into the way your team already works.
- Web applications Custom browser-based software for the work your team does every day, from booking tools to operations systems.
- Dashboards and reporting Dashboards that pull numbers from the systems you already use and show what matters without a spreadsheet.
- Multi-tenant architecture One product serving many customer accounts, with each account's data kept separate and secure.
Solutions
View all solutions- Automated reporting Reports that build themselves from your systems on schedule, with a plain-language summary of what changed.
- Knowledge assistant (RAG) Ask a question in plain words and get an answer from your own documents, with links to the sources.
- Forecasting dashboards Forecast demand from your sales history and seasonality, and flag items to reorder before they run out.
- Internal help desk assistant An assistant in Slack or Teams that answers policy and how-to questions from your handbooks and opens tickets when needed.
- Payroll automation Turn pay rules, attendance and leave into tested software so the monthly run becomes a review and payslips go out in one batch.
- AI sales research agent Short, sourced briefings on each prospect before a call, drafted by an AI agent from public information and your CRM.
Industries
View all industriesCase studies
View all case studiesGuides and articles
View all guides and articles- How to connect AI to your database safely Let staff ask questions of your data in plain English without risking production systems or sensitive records.
- PostgreSQL vs MySQL The two most popular open-source databases compared for business applications.
- What is RAG and when do you need it? How retrieval-augmented generation lets AI answer from your documents, when it fits and how to build it well.
- How to automate reporting with AI and MCP Build reports that assemble themselves from your systems through MCP, with a written summary people can trust.
- How to set up automated reporting Replace manual weekly and monthly reports with dashboards and summaries that update themselves.
- RAG vs fine-tuning Two ways to make AI work with your knowledge, compared by cost, accuracy and upkeep.
AI models
View all ai models- text-embedding-3-large OpenAI's most capable embedding model for search and retrieval, with 3,072-dimension vectors and support for English and other languages.
- text-embedding-3-small OpenAI's efficient, lowest-cost embedding model, with 1,536-dimension vectors for search, retrieval and similarity at scale.
- Cohere Embed v4 Cohere's multimodal embedding model for enterprise search, embedding text, images and mixed documents, with flexible vector sizes and long inputs.
- Voyage 4 Voyage AI's current embedding series, now part of MongoDB, with large, standard and lite models that share one embedding space.
- Embeddings Models that turn text into vectors for search, retrieval, clustering and recommendations.
- Gemini Embedding 2 Google's current embedding model, multimodal: it embeds text, images, video, audio and PDFs for search and retrieval.
Glossary terms
View all glossary terms- RAG RAG, or retrieval-augmented generation, is a method where an AI system first finds relevant passages in your documents, then answers using only those passages.
- Database A database is an organized store of data that software can search, update and keep consistent, such as customers, orders or bookings.
- Vector database A vector database stores embeddings, lists of numbers that represent meaning, and quickly finds the ones most similar to a query.
- Embeddings Embeddings are lists of numbers that represent the meaning of text or images, so software can find items that are similar in meaning.
- Semantic search Semantic search finds results by meaning rather than exact words, so a search for "reset my login" can find an article titled "Forgot your password".
- Backend The backend is the part of an application that runs on servers, storing data, applying business rules and serving the frontend through APIs.
FAQ
Questions people ask us
Have a question that is not here? Ask us directly.
Yes. It is maintained by Chroma.
Not for local use. Chroma Cloud needs a tenant, database and API key.
Yes. It can delete documents and whole collections, so review changes.