# Deep Lake

Deeplake is AI Data Runtime for Agents. It provides serverless postgres with a multimodal datalake, enabling scalable retrieval and training.

Repository: https://github.com/activeloopai/deeplake
Canonical: https://ross.abutalabs.com/products/deep-lake
Homepage: https://deeplake.ai
Language: C++
License: Apache-2.0
License Family: permissive
Topics: deep-learning, pytorch, ai, mlops, computer-vision, datalake, llm, large-language-models, vector-database, postgres, agent, agentic-rag, clawbot, filesystem, memory, multimodal, openclaw, rag, skill
Last push: 2026-05-21T15:28:00+00:00
Link (homepage): https://deeplake.ai

## Health v2 (maintenance only)
Score: 77/100 (v2, computed 2026-09-03T02:39:23.370411+00:00)
- activity 83, release rhythm 57, longevity 100
- inputs: {"age_days": 2581, "days_push": 104, "days_rel": 204, "gap_med": 5.0, "n_releases_24m": 63}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 9228, forks 723 (observed 2026-08-28T04:10:28.981892+00:00)

## What it is
Deep Lake is an open-source database for AI that stores multimodal data (images, video, audio, text, embeddings, annotations) in a format optimized for deep learning and LLM applications. It provides vector search, data streaming for training at scale, dataset versioning, and a serverless Postgres-compatible interface for building agentic RAG systems.

## Use cases
- store and search embeddings for building LLM applications
- manage large multimodal datasets while training deep learning models
- stream training data from S3 to PyTorch or TensorFlow at scale
- build agentic RAG pipelines with long-term memory for agents
- version and track lineage of AI datasets
- run vector search over images, video, audio, and PDFs

## When to choose
- you need a multimodal datalake plus vector search in one system
- you train deep learning models on large datasets stored in the cloud
- you are building LLM or agent applications that need retrieval over diverse data types
- you want dataset versioning and lineage for ML workflows

## When to avoid
- you only need a general-purpose relational database for transactional workloads
- you need a lightweight single-purpose vector store with minimal dependencies
- your data is small and fits comfortably in a traditional SQL database

## Facets
- artifact type: library
- maturity: active
- function: vector-database, database, rag, machine-learning, deep-learning, data-science, etl
- domain: databases, machine-learning, deep-learning, large-language-models, computer-vision
- platform: python, cross-platform, cloud
- tags: datalake, multimodal, vector-search, mlops, pytorch, data-versioning, serverless, embeddings, llm-memory, agentic-rag, retrieval-augmented-generation, data-engineering, ai-agents, gpu

## Member repositories
- activeloopai/deeplake (main) score 77
- activeloopai/hivemind (sdk) score 76

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:10:28.981892+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-29T17:23:15.891396+00:00, confidence not recorded.
  - readme: https://github.com/activeloopai/deeplake (fetched 2026-08-28T04:10:28.981892+00:00, sha db738d04e2f9)
  - homepage: https://deeplake.ai (fetched 2026-08-29T08:23:26.494408+00:00, sha 92d15715bf8d)
- Data as of 2026-08-30T08:39:29.467469+00:00.