Chroma Framework Integrations & SDKs 2026
Chroma ships native integrations for LangChain, LlamaIndex and Haystack, with first-party Python and JavaScript SDKs. No official Ruby, PHP or Go client.
Chroma Framework Integrations verdict
Chroma ships native integrations for LangChain, LlamaIndex, Haystack, Google ADK, VoltAgent and Mem0.
Plus observability connectors for DeepEval, Braintrust, OpenLLMetry and Anthropic MCP. First-party SDKs cover Python, at v1.5.9, and JavaScript or TypeScript, at v3.4.3.
Pick Chroma if your stack is Python or JavaScript on LangChain or LlamaIndex. Setup takes a few lines, with no credentials required for local use. If you need an official Go, Ruby or PHP client, or a framework not on Chroma's integration list, look at alternatives.
- There is no official Ruby, PHP or Go SDK. Those are community-maintained and unsupported by Chroma directly.
- The JavaScript client v3, the chromadb npm package, is only compatible with the chromadb Python server v1.0.6 or newer, or Chroma Cloud. Older servers require JS client v2.
- The Haystack and VoltAgent integrations are single-language, Python-only and JS-only respectively, not cross-language.
- Native RAG frameworks
- 3 (LangChain, LlamaIndex, Haystack)
- Official SDK languages
- 2 (Python, JS/TS)
- Python SDK version
- v1.5.9 (May 2026)
- JS/TS SDK version
- v3.4.3
- Observability tools
- 3 (DeepEval, Braintrust, OpenLLMetry)
Framework and observability integrations
| Integration | Type | Capabilities | Setup |
|---|---|---|---|
| LangChain | Native | RAG orchestration · Similarity search · MMR retrieval | Low |
| LlamaIndex | Native | Vector index · Metadata filtering · Storage context | Low |
| Haystack | Native | Search pipeline · Document retrieval | Low |
| Google ADK | Native | Agent orchestration · Tool use | Medium |
| VoltAgent | Native | TypeScript agents · RAG tool | Medium |
| Mem0 | Native | Long-term memory · Agent memory layer | Low |
| DeepEval | Observability | RAG evaluation · LLM testing | Low |
| Braintrust | Observability | Eval logging · Tracing | Low |
| OpenLLMetry | Observability | OpenTelemetry tracing · LLM observability | Low |
| Anthropic MCP | Protocol | Claude Desktop integration · MCP server | Medium |
Find your integration path into Chroma
- LangChain is a native integration (RAG orchestration, Similarity search, MMR retrieval), low setup effort.
- Official Python SDK v1.5.9 targets API .
- Install:
pip install langchain-chroma
Based on Chroma's documented integrations and official SDK matrix.
LangChain + Chroma quickstart
from langchain_chroma import Chroma
from langchain_openai import OpenAIEmbeddings
# Create or load a persistent Chroma collection via LangChain
vector_store = Chroma(
collection_name="my_collection",
embedding_function=OpenAIEmbeddings(),
persist_directory="./chroma_db",
)
# Index documents
vector_store.add_documents(docs)
# Retrieve by similarity with a metadata filter
hits = vector_store.similarity_search("how does billing work?", k=4, filter={"source": "docs"})
# Or use Maximum Marginal Relevance to reduce redundancy
retriever = vector_store.as_retriever(
search_type="mmr", search_kwargs={"k": 4, "fetch_k": 10}
)Official SDKs and language clients
| Language | Package | Version | Status |
|---|---|---|---|
| Python | chromadb | v1.5.9 | Official |
| JavaScript/TypeScript | chromadb (npm) | v3.4.3 | Official |
| Python ≥3.9 | chromadb | requires Python ≥3.9 | Requirement |
| Rust | chromadb-rs (community) | — | Community |
Chroma LangChain integration detail
- Install via pip install langchain-chroma>=0.1.2
- Exposes the Chroma class from langchain_chroma.vectorstores
- Supports similarity_search with k and filter keyword arguments
- Supports similarity_search_with_score for distance-ranked results
- MMR (Maximum Marginal Relevance) retrieval available via as_retriever with search_type=mmr
- No credentials needed; works with a local Chroma instance out of the box
Chroma LlamaIndex integration detail
- Install via pip install llama-index-vector-stores-chroma (latest v0.5.5, Dec 2025)
- Exposes ChromaVectorStore class from llama_index.vector_stores.chroma
- VectorStoreIndex wraps ChromaVectorStore via StorageContext.from_defaults
- Supports metadata filter operators: exact match, FilterOperator.GT (greater than), AND/OR logic
- Compatible with both EphemeralClient (in-memory) and PersistentClient (disk) Chroma modes
- llama-index-vector-stores-chroma is listed as a community/integration package on PyPI
Chroma Framework Integrations FAQ
Which AI frameworks integrate natively with Chroma?
LangChain in Python and JavaScript, LlamaIndex in Python and JavaScript, Haystack in Python only, Google ADK in Python and JavaScript, VoltAgent in JavaScript only, and Mem0 in Python only. Chroma also integrates with the evaluation tools DeepEval, Braintrust and OpenLLMetry.
What languages does Chroma have official SDK support for?
Python, through chromadb v1.5.9, released May 2026, and JavaScript or TypeScript, through chromadb npm v3.4.3. Rust support exists as a community client. Ruby and PHP are not officially supported by Chroma.
How do I install and use Chroma with LangChain?
Install with pip install langchain-chroma>=0.1.2 and chromadb. Import Chroma from langchain_chroma, point it at a collection name and embedding function, call add_documents to index, then use similarity_search with a query, k=4 and a filter to retrieve. MMR retrieval is available through as_retriever with search_type='mmr'.
How do I use Chroma with LlamaIndex?
Install llama-index-vector-stores-chroma, at v0.5.5 from December 2025, and chromadb. Import ChromaVectorStore from llama_index.vector_stores.chroma, wrap it in StorageContext.from_defaults, then build a VectorStoreIndex. Metadata filtering supports exact match, greater-than, and AND or OR operators.
Is the JavaScript Chroma client compatible with all chromadb server versions?
No. The current npm package chromadb v3.x is only compatible with the chromadb Python server v1.0.6 or newer, or Chroma Cloud. If you are running an older server, you need JS client v2. Embedding functions are no longer bundled in v3 and must be installed separately through @chroma-core/default-embed.
Does Chroma integrate with LLM observability tools?
Yes. Chroma has native integrations with DeepEval, in Python for RAG evaluation and LLM testing, Braintrust, in Python and JavaScript for eval logging and tracing, and OpenLLMetry, in Python with JavaScript support coming. These let you instrument RAG pipelines without extra glue code.
Sources & verification
| Source | What was checked | Last checked |
|---|---|---|
| Trychroma Official | Official product page | July 10, 2026 |
| Langchain Developer docs | Independent reference | July 10, 2026 |
| Llamaindex Vector Stores Chroma Metadata Filter | Independent reference | July 10, 2026 |
| Pypi Project Chromadb | Project Chromadb | July 10, 2026 |
| Trychroma Developer docs | Integrations Chroma Integrations | July 10, 2026 |
Every fact on this Chroma page is tied to a named source and a verification date. Freshness-sensitive figures trace to the sources above; verify against the vendor before relying on them.
Explore Chroma
Every page on Chroma in one place, you are on framework integrations.