Product updates

Recent improvements to the platform, modules, remote support, and customer-facing operations.

2026-08-28

The Customer Portal suite is available on Standard plans and above. Brand a portal, invite customers, and design the device UI they see — one portal and one customer on Standard; unlimited on Business and above.
Operators get a map, alerts, and cameras in the portal. Find devices, acknowledge product alerts, test connectivity, and review live video with seekable event history.
Starter kits jump-start common connected products. Apply ready-made patterns for silo outload, energy sites, HVAC comfort, industrial gateways, remote signage, and camera access.
Class software and UI design ship with the suite. Lay out the OEM device experience and manage class containers and rollouts alongside the portal.

Product Applications are available when enabled for the organisation. Author the product your customers run — versioned models, Device Class bindings, rules, operations, and portal instances — then publish updates without rebuilding the portal from scratch.

Fleet composition shows how a class is really made up. See distributions across agent version, OS, packages, and more, then click through into the matching device cohort.
Makeup history sits alongside live Pulse. Compare how class composition has shifted over time, not only the current snapshot.

2026-08-24

Video transfer, storage, and transcode appear on Usage and pricing. Plan limits and metered video costs are visible in-product instead of only in support conversations.

2026-08-22

Run bounded investigations from device detail. Capture host logs for a short window, refresh inventory facts, or continue with connection traces and Device Check from the traces panel.

2026-08-21

Device classes support dual-camera widgets. Bind a second camera, name cameras per device, and go live over HLS.
Level inputs are first-class on the Device Class. Define composed level inputs and show derived level tiles to operators — built for tanks, silos, and similar hardware.

2026-08-20

Project Boards are available in the core portal when enabled for the organisation. Organise delivery work as Kanban boards in Dataplicity so linked cards stay one click from devices, events, incidents, and monitors.
Add work from fleet surfaces. Create cards from device, incident, and event views; use checklists, labels, attachments, and archiving.
Private boards use matching labels and access tags. Keep sensitive triage restricted to groups whose board access tags match a board label, while initiatives and projects roll up from linked card progress.

2026-08-11

Channel samples no longer land in Postgres. Agents still upload a combined batch; last-known and the recent window stay on Redis, and history is persisted to object storage.
Retention follows the plan. Standard keeps 7 days, Pro keeps 90 days, and Enterprise can enable 365-day extended retention (beta, enterprise only; may be priced after GA).
Channel history is defined on the class stream. Devices receive durability with their rate limits and keep that contract across restarts. Live-only channels stay last-known and can sample faster; keep-history channels use the plan window.
Overloaded ingest publishes coverage instead of going dark. Limits include a `coverage` factor; agents downsample to it, and the server thins still-entitled uploads rather than 429ing the fleet to zero.

2026-08-01

Device Class Pulse is available on paid plans. Dataplicity learns the normal operational fingerprint of each Device Class, then surfaces devices that drift — disk, packages, agent version, and more — with plain-language findings and charts.

SIM cards are managed in one inventory. Sync Hologram or Jasper, or add SIMs manually, assign them to devices, and manage provider setup from Integrations (Business and above).
Chat integrations live under Integrations. Legacy Slack webhooks move into the shared hub alongside Teams; delivery outcomes show up in Logging under the Integrations source.

Hosted MCP is live at mcp.dataplicity.com. Connect Cursor (and other MCP clients) with an organisation API key; Gateway also expands fleet jobs, monitors, wormhole, and timeline coverage.
Gateway docs point at the canonical documentation site. The developers landing page is clearer, with OpenAPI still available for machines.

Paid plans can start Human Support from the assistant chooser. Prefer an AI assistant first, or go straight to the human team when you need one.
Paste screenshots and logs into AI chats more reliably. Suricata and other attachment-capable assistants accept real image and log pastes.

Engineering device detail is built for observe-and-manage. Overview puts connection quality, wormhole, monitors, reachability, and disk-space history up front; Terminal stays the place to act.
Connection traces show live progress. Watch collection finish instead of refreshing to see whether a trace completed.

Public status pages show the customer-facing incident note. Internal create/publish/paging noise stays out of the public timeline.
Incidents, alerts, and on-call escalations are on Business plans (and above). The ops suite is out of beta.

2026-07-27

Device Check is free and public. Run one command on a Linux device (no account required) and stream health, network-path breaks, clock skew, and process offenders into a temporary shareable report.

File requests are available on Business plans and above. Create a request, share a link or curl command, and collect files from the field without a full device session.
Business plans can edit device tags from the remote device UI.

Incidents are built for triage. Ownership, severity, and urgency are front and center so open incidents drive response.
Named escalation paths route on-call. Group schedules and notification channels, assign a path on an active incident, and see who’s on duty by path on summaries and dashboards.
Get emergency pages in the Dataplicity app. Critical alerts can reach you from monitoring rules you configure, or when someone raises an incident manually.

Web console supports multiple sessions. Keep several device consoles open in tabs and switch without losing state.
Monitor history is easier to scan. Endpoint, CORS, DNS, heartbeat, and fleet monitors show recent evaluation history inline; User Impact rows include short per-flow sparklines.
You can keep talking to AI while Human Support is pending. Continue with the assistant instead of waiting idle for a handoff.