LoginSign Up
← Back to all sparks
E

Encord

DEVOPS
Velocity2.5

Encord pushes labeling toward agentic, multi-file workflows.

data-labelingai-agentsworkflowsconsensus-reviewcomputer-visiondataset-management
◆Current state
Encord is making its labeling pipeline more automated and more complex — agents from the catalog can now be added as workflow nodes, multi-file Data Groups went GA, and Labels in Index went GA across all datasets. UX and integrity work — consensus-review username hiding, a metadata panel, webhook signature verification — round out the recent shipping.
◆Where it's heading
The product is splitting into two layers: an automation runtime where AI agents handle parts of labeling pipelines without manual triggers, and a richer data plane where multi-file groupings, label exploration, and consensus review are first-class objects. Encord is packaging more of the labeling-ops workflow into the platform rather than leaving it to custom integration code.
◆Prediction
Expect the Agents Catalog to expand with pre-built agents for common pre-labeling and QA tasks, and expect Index to keep absorbing labeling-aware exploration features now that labels are exposed there.

◆Recent moves

  1. 24d ago

    Hide annotator usernames in Consensus Review to prevent reviewer bias

    Reviewers can no longer see which annotator made a label, removing a known source of consensus-review bias. Small but the kind of integrity feature buyers in regulated domains specifically ask about.

    View source ↗
  2. 1mo ago

    Agents become first-class Workflow nodes; webhook signing ships

    ⚡ SPARK

    Agents move from manual editor tools to first-class Workflow nodes, running directly from the Project Queue. Combined with webhook signature verification, the release reads as production-hardening for unattended labeling pipelines.

    View source ↗
  3. 1mo ago

    ​Metadata Panel for Taskers

    A dedicated in-editor panel surfaces task metadata with copy and link interactions, updating per active tile. Quality-of-life for taskers working in Data Groups where context lives outside the active media.

    View source ↗
  4. 1mo ago

    Labels in Index (General Release)

    Label-based exploration across full datasets becomes generally available, with visual overlays of bounding boxes and polygons for both ground truth and model predictions. Pushes label data into the same surface used for dataset curation.

    View source ↗
  5. 2mo ago

    Improved review UX for consensus classifications

    Refines consensus classification review with by-annotator and by-label toggles plus expand-and-collapse. Pairs with the username-hiding update to make consensus review both more usable and less biased.

    View source ↗
  6. 2mo ago

    Data Groups (General Release)

    Multi-file labeling tasks graduate to general availability with cross-tile instances, global classifications, and flexible tile management. Mature support for mixed-media projects (medical imaging, multi-camera) that do not fit single-file pipelines.

    View source ↗