Skip to main content
Pixeltable Cloud extends the local SDK with team collaboration and production deployment capabilities.

​
Publish & Replicate ✅ Available Now

FeatureAPI
Publish datasetspxt.publish(source, destination_uri)
Replicate datasetspxt.replicate(remote_uri, local_path)
Sync updatespush(), pull()
Access controlaccess='public' or 'private'
# Publish your curated dataset
pxt.publish(source='my-table', destination_uri='pxt://myorg/my-dataset')

# Anyone can replicate public datasets (no account required)
coco = pxt.replicate(remote_uri='pxt://pixeltable:fiftyone/coco_mini_2017', local_path='coco-copy')

Data Sharing Guide

Full documentation on publish, replicate, push, and pull

​
2. Endpoints 🔜 Coming Soon

One-command API deployment with managed hosting.
FeatureWhat
pxt serveLocal development server
pxt deployCloud deployment with auto-scaling
Pre-signed URLsMedia access without proxying
Join the waitlist to get early access to Endpoints.

​
3. Live Tables 🔜 Coming Soon

Multi-writer collaboration and serverless compute.
FeatureWhat
Multi-writerTeam collaboration on shared tables
Serverless computeAuto-scaling without infrastructure
UDF versioningSafe experimentation with code changes
RBAC + auditGovernance and compliance

​
Unified: Where It’s All Going

When all three cloud services are available, the two use cases converge:
  • Data wrangling + AI pipelines + endpoints = one system
  • Orchestration + storage + retrieval unified
  • Table becomes the endpoint
Your training datasets and production APIs share the same infrastructure—versioning, lineage, and retrieval in the serving path.

Book a Demo

Schedule time with our team to discuss your use case
Last modified on January 25, 2026