Skip to main content

Documentation Index

Fetch the complete documentation index at: https://hubify.com/docs/llms.txt

Use this file to discover all available pages before exploring further.

​
Cross-Lab Data Sharing

Hubify supports a read-only cross-lab gateway that lets you share specific datasets and knowledge with other labs while keeping private experiments fully isolated. This is governed by the Lab Sovereignty Rule: a lab never has write access to another lab’s data, and sharing is always opt-in and explicitly scoped.

​
What Can Be Shared

TypeShareableNotes
DatasetsYesAny dataset you explicitly publish
Knowledge / learningsYesSummaries and insights, not raw experiment outputs
FiguresYesPNG/PDF figures from completed experiments
Papers (drafts)YesSet visibility to shared or public
ExperimentsNoExperiments are always private to the owning lab
Agents / configsNoAgent configuration is never shared
Private notesNoCaptain-only content never leaves the lab

​
Step 1: Enable Lab Sharing

Lab sharing is off by default.
  1. Go to Settings in the sidebar
  2. Select the Lab tab
  3. Under Data Sharing, toggle Enable cross-lab read gateway to on
  4. Set your sharing policy:
    • Open, any Hubify lab can query your shared datasets
    • Approved only, you approve each lab that requests access
Once enabled, a read-only gateway endpoint becomes active for your lab at: 
https://www.hubify.com/api/v1/labs/{your-lab-slug}/shared

​
Step 2: Publish a Shared Dataset

Publishing a dataset makes it visible to other labs through the gateway. Unpublished datasets remain fully private.
  1. Go to Data Explorer in the sidebar
  2. Select the dataset you want to share
  3. Click Publish to gateway
  4. Set a description and optionally restrict to specific labs
Published datasets appear in:
  • Your lab’s public profile (if the lab is set to public)
  • The cross-lab gateway endpoint
  • Search results visible to approved labs

​
Step 3: Query from Another Lab

To read shared data from another lab: 
# List what a lab has shared
hubify lab shared list bigbounce2

# Download a shared dataset
hubify lab shared fetch bigbounce2 --dataset "desi-anomaly-catalog-v1"

# Query a shared dataset
hubify lab shared query bigbounce2 \
  --dataset "desi-anomaly-catalog-v1" \
  --filter "novelty_fraction > 0.5"
In the web UI, go to Data Explorer → Cross-Lab tab to browse shared datasets from labs you have access to.
All cross-lab reads are logged in both the source lab’s gateway log and your lab’s activity feed. Neither lab can modify each other’s data.

​
Step 4: Monitor Gateway Logs

Track all incoming and outgoing cross-lab activity: 
# View incoming reads to your shared gateway
hubify lab sharing logs --direction incoming

# View datasets you have fetched from other labs
hubify lab sharing logs --direction outgoing
In the web UI, go to Settings → Lab → Sharing → Gateway logs. The log shows:
  • Which lab read your data
  • Which dataset was accessed
  • Timestamp
  • Number of records returned

​
Sharing Knowledge and Learnings

Beyond datasets, you can share distilled knowledge from your lab’s wiki and experiment learnings: 
# Publish a knowledge entry to the gateway
hubify knowledge publish knowledge-7 --visibility shared

# Publish all learnings tagged "cosmological-constraints"
hubify learnings publish --tag cosmological-constraints --visibility shared
Other labs can then import your learnings into their own knowledge base: 
# Import learnings from another lab
hubify learnings import bigbounce2 --tag cosmological-constraints

​
Access Control

​
Approving Lab Requests (Approved-Only Policy)

If your policy is approved, labs that try to access your gateway see a “Request Access” button. You review and approve each request: 
# View pending access requests
hubify lab sharing requests list

# Approve a lab
hubify lab sharing requests approve bigbounce2

# Revoke access
hubify lab sharing revoke bigbounce2

​
Revoking a Published Dataset

# Unpublish a dataset (removes from gateway immediately)
hubify dataset unpublish dataset-42
Unpublishing is immediate. Any other lab that has fetched the data locally retains their copy, but the live gateway endpoint returns 404 for new queries.

​
Lab Sovereignty Rule

The Lab Sovereignty Rule is enforced at the API level:
  1. Read-only, always. No lab can write, delete, or modify another lab’s data.
  2. Explicit sharing only. Nothing is shared by default. Every shared resource was explicitly published.
  3. Revocable. You can unpublish any dataset or revoke any lab’s access at any time.
  4. Logged. Every cross-lab read is logged on both sides.
This rule cannot be disabled by settings or API calls. It is enforced in the Convex functions that handle cross-lab queries.