Features

One workflow, end to end

MASS isn't a pile of disconnected tools. The agent, the labeler, and the participation map are stages of a single pipeline: describe what you're studying, capture it on the video, and see the result as data you can explore and share.

Agent-guided workflows

Atlas: An AI agent that runs the workflow with you

Atlas is a conversational assistant built into MASS. Instead of clicking through menus, you tell it what you're trying to do, and it does the setup, organization, and first-pass analysis for you.

Analyzes video for you: Describe the activities you care about ("students raising hands", "group discussion") and Atlas scans the footage and proposes a full annotation pass
Manages your workspace: Creates projects and datasets, attaches files, and keeps your study organized through plain conversation
Never acts without you: Anything that creates or changes data is approval-gated; Atlas proposes, you confirm
Fully auditable: Every action the agent takes is logged, so your methodology stays defensible

ATLAS

Find every segment in this week's recordings where students are working in groups.

I analyzed 12 videos in Dataset: Week 14 and found 47 candidate segments of group work. I'd like to create an annotation pass called "group-work" so you can review them in the labeler.

Create annotation pass "group-work"?

Annotation

The Labeler: Capture exactly what's happening

The labeler is where observation becomes data. Annotate any subject in the frame and track them across the whole video: precisely, quickly, and together with your team.

Bounding boxes with keyframes: Mark a subject at a few moments and MASS interpolates their position between them
AI click-to-segment: Click a subject once and get a pixel-accurate outline that tracks them through the video, refined with simple include/exclude clicks
Multi-subject, multi-pass: Code several students and several behavior categories in the same session, in separate organized passes
True team coding: Colleagues annotate the same video simultaneously; per-subject locks and automatic merging keep everyone's work intact
Review workflow built in: Labelers submit work for review; reviewers approve or request changes with notes, just like second-coder practice

LABELER — session_04.mp4 · pass: group-work

Student A · on-task

Student B · discussing

Visualization

The Participation Map: See the whole session at once

Every annotation becomes a colored segment on an interactive, frame-accurate timeline. What took hours of footage to record takes seconds to read.

Group your view: Lanes by student, by activity, or by coding pass; the same data, three perspectives
Filter to the question: Narrow to specific subjects, activities, or tags and the timeline updates live
Linked to the video: Click a segment to jump straight to that moment; the playhead tracks playback on the map
Saved views & sharing: Save a filter-and-grouping setup as a named view, and share a read-only link with collaborators or stakeholders
Notes where they belong: Attach observations to specific segments instead of a separate document

PARTICIPATION MAP — grouped by activity

On-task

Discussing

Hand raised

Transitions

00:0015:0030:0045:00

Teamwork & trust

Built for teams and for rigor

Research coding is a team sport, and methodology matters. MASS treats both as first-class concerns rather than afterthoughts.

Roles that match your team: Viewers, Editors, Reviewers, and Managers per dataset; invite collaborators with exactly the access they need
Conflict-free co-editing: Subject-level locks mean two coders never silently overwrite each other; everyone's edits merge in real time
Checkpointed history: Committing annotations snapshots a revision while leaving the draft open, so you always know which version a finding came from
Notifications that keep work moving: Reviewers and labelers are notified in-app and by email when something needs their attention
Secure by design: Scoped authentication, per-resource permission checks, and audited AI actions throughout the platform

Submit

A labeler finishes a pass and marks it ready for review.

Review

The assigned reviewer inspects the work and requests changes, with a reason the labeler sees in context.

Approve

Approval commits the annotations as a checkpointed revision, ready for analysis.

Want a walkthrough?

We'll demo the full pipeline — agent, labeler, participation map — on footage like yours.

Request a demo