Skip to main content

Getting Started

Interacting with Mosaico requires two components: mosaicod, which runs server-side and manages all data operations, and a client SDK to communicate with it from your code.

Mosaico Daemon​

mosaicod is the engine behind the platform. It handles storage, catalog management, ingestion, and retrieval. No client call will succeed without a running instance.

Use the folling compose file for a quick local setup:

compose.yaml
services:
  db:
    image: postgres:18
    environment:
      POSTGRES_HOST_AUTH_METHOD: trust

  mosaicod:
    image: ghcr.io/mosaico-labs/mosaicod:latest
    environment:
      MOSAICOD_DB_URL: postgresql://postgres@db:5432/postgres
      MOSAICOD_STORE_ENDPOINT: file:///tmp
      MOSAICOD_STORE_BUCKET: mosaico
    command: run --host 0.0.0.0
    depends_on:
      - db
    ports:
      - "6726:6726"

This setup will create a data folder at /tmp/mosaico on your machine. You can change this by modifying the MOSAICOD_STORE_ENDPOINT variable.

note

For more advanced installation options, see the installation guide.

Python SDK​

The mosaicolabs Python SDK is the primary way to interact with the platform. It provides a high-level API for the full data lifecycle, ingesting sensor data, querying catalogs, and streaming data into ML pipelines, without any custom serialization code.

With mosaicod running, the SDK is all you need to start working with your data. Install it, point it at your daemon, and you have full programmatic access to the platform: write sequences, query catalogs, and pull data directly into your pipelines. No extra configuration or intermediary services required.

Install it via pip:

pip install mosaicolabs

The following example connects to a local daemon instance and lists available sequences:

lets_get_physical.py
from mosaicolabs import MosaicoClient

# Connect to the Mosaico server
with MosaicoClient.connect(host="localhost", port=6726) as client:
    # List available sequences
    sequences = client.list_sequences()
    print(f"Connected! Found sequences: {sequences}")

Jump straight into the how-to guides on writing data and querying sequences, or see the Python SDK documentation for the full reference.