Morphik Core is an open-source multimodal retrieval engine by Morphik for AI back-office workflows. It offers ingestion, search (multimodal), and metadata extraction with Python SDK and REST API. The project has a large community and active development as of 2026-07-23.
Collecting history — the radar snapshots this repo daily. The trend line appears after 3 days of data (1 so far).
What it is
Morphik Core is an open-source multimodal retrieval engine that ingests, searches, and manages unstructured and multimodal documents. It provides tools for multimodal search and fast metadata extraction, with integrations and a Python SDK plus a REST API.
How it works
The README describes ingesting files via the Python SDK and querying data through a simple interface. It emphasizes multimodal search using ColPali to understand visual content and supports extraction of bounding boxes, labeling, and classification. It also mentions an on-prem self-host option and a web-based Morphik Console for ingesting, searching, and querying data.
Getting started
- Code example for ingestion:
from morphik import Morphik
morphik = Morphik("<your-morphik-uri>")
morphik.ingest_file("path/to/your/super/complex/file.pdf")
- Code example for querying:
morphik.query("What's the height of screw 14-A in the chair assembly instructions?")
- Documentation links include Python SDK and REST API references, and a Morphik Console for web interaction.
Recent releases
No releases are listed in the provided data (RELEASES (latest 0): - none).
Traction
Stars: 3677. Forks: 316. Open issues: 60. Language: Python. License: none listed. Created: 2024-11-11. Last push: 2026-07-23.
Behind the repo
The project is part of Morphik, offering AI back-office tools for skilled nursing and senior living operators. It provides an open-source core behind paid products and a dev portal at dev.morphik.ai.
Caveats
License section states: Morphik Core is source-available under the Business Source License 1.1. Licensing terms include free personal/indie use, free commercial production under certain revenue thresholds, and future re-licensing to Apache 2.0 after four years from first release. No other license details are provided in the excerpt.