FAQ

Getting started

Do I need an account to try Connectify?

No. Click Start building on the landing page and you're in. Your work persists to local storage. Create an account when you want to sync across devices, share with teammates, or publish to the community.

Which browsers are supported?

Chrome, Firefox, Safari, and Edge — current and one previous major version. We test most actively on Chrome. Mobile browsers work for read-only views but the editor expects a pointer.

Does Connectify work offline?

The editor works offline — graphs persist to local storage and reopen without a connection. Sharing, presence, and the Community tab require a connection.

Storage & data

Where is my work saved?

By default, in your browser's local storage. Signed-in users sync to the cloud automatically. Local-only projects show a small "local" chip in the Hub.

What's the storage limit?

Local storage caps at ~5 MB per origin in most browsers. That's enough for hundreds of typical graphs. History keeps the last 50 snapshots per variant; older states drop silently. If you need long-term checkpoints, fork the graph or sign in to enable cloud sync.

Can I export a graph?

Yes. From the editor, File → Export writes a portable JSON. Import it on another machine or share the file directly. Run data is included optionally.

Can I import an existing notebook or pipeline?

We're working on importers for Jupyter notebooks, Airflow DAGs, and Hugging Face pipelines. Reach out to support@connectifyai.com if you have a specific format you'd like supported.

Collaboration

How do I share a graph with someone who isn't on my team?

Use Share → Viewer link. The link grants read-only View Mode access without requiring login. Revoke any time from the same dialog.

What happens if two people edit at the same time?

Connectify uses last-writer-wins for positions and config values, and operational merging for connections. In practice, conflicts are rare. If you need true branching, fork the graph and merge back later.

Is there an audit log?

Version history shows every change with author and timestamp. Enterprise workspaces get a separate audit log for permission changes, exports, and publishes.

Performance & limits

How large a graph can the editor handle?

The canvas is comfortable with a few hundred nodes. Beyond ~500 nodes, group related nodes into subgraphs to keep the canvas responsive.

How long does a run take?

Light graphs (preprocessing, small inferences) run in seconds entirely in the browser. Heavier graphs (training, large model inference) offload to a runtime worker; expect minutes to hours depending on the workload.

Can I run on my own hardware?

Yes — self-hosted runtimes are supported. Configure a runtime URL in your workspace settings and the editor offloads runs to it. Useful for proprietary models, large datasets, or air-gapped environments.

Publishing & community

How do I publish a graph?

From the editor, click Publish. Provide title, description, tags, license, and visibility. Public graphs appear in the Community tab within minutes.

Can I unpublish?

Yes, any time, from the same dialog. Existing forks remain (they're independent copies), but the original is removed from the catalog.

What about licensing?

Pick a license when publishing. The catalog supports MIT, Apache 2.0, CC-BY, CC-BY-SA, and a custom-text option. Forks inherit the source's license obligations.

Errors & troubleshooting

"This connection isn't compatible"

The two port types don't match and no adaptor exists. Options: insert a Logic transform that bridges them, write a custom adaptor (see Type System & Adaptors), or change the source/destination node.

"Run failed at node X"

Open the failing node and check its Run Data tab — the error message and stack trace appear there. Most failures are config issues (wrong path, missing API key) or upstream data shape mismatches.

"Local storage is full"

Browser storage is capped. Either delete old projects from the Hub, export them as JSON for cold storage, or sign in to enable cloud sync.

The canvas froze.

Refresh the page. Your work is persisted; you won't lose anything. If it happens repeatedly, your graph may be too large — consider grouping into subgraphs or splitting into separate projects.