Seqvio

Open source · Local · Agent-driven

Seqvio

An explainer video toolkit for agents

Turn real terminal sessions, browser runs, and authored ideas into reviewable videos where narration, visuals, and evidence move together.

Watch the product demo Install the skill
Real evidence Phrase-anchored timing Deterministic QA
See how it works

One connected workflow

From real agent work to an explanation

Seqvio keeps the source activity, editorial decisions, spoken phrases, and final frames connected instead of reconstructing the story after the fact.

  1. 01

    Capture real evidence

    Preserve terminal and browser activity in the order it happened.

  2. 02

    Review the story

    Make the claim, evidence order, and visual direction inspectable.

  3. 03

    Align voice and visuals

    Bind exact spoken phrases to actions, targets, and evidence.

  4. 04

    Validate and render

    Catch blank frames, overflow, pacing, and broken references.

Joint explanation timing

The voice sets the clock

Seqvio resolves synthesized speech first, then maps each visual action to the phrase it explains. Authored and captured order stays intact without hand-tuned padding.

Exact anchors Spoken phrases, not guessed timestamps Semantic order Visual steps keep their intended sequence One contract Narration, targets, actions, and evidence
Measured speechresolved after TTS
capture the run1.42s
review the story1.58s
show the evidence1.36s
Visual sequencesame semantic order
01Terminal evidencesource
02Story structureaction
03Evidence focustarget

Product Hunt demo · 72 seconds

See the complete system in motion

Follow one real skill evaluation from an agent session to browser evidence, story review, phrase-level timing, QA, and reusable explanation formats.

One system, many explanations

Use the visual form the claim needs

The workflow stays consistent while the evidence and visual structure change for each explanation job.

01

PR video review

Turn a change, its evidence, and the review conclusion into a concise walkthrough.

02

Tool comparison

Compare products across the criteria that matter, then make the trade-off visible.

03

Tutorial verification

Walk through each instruction and show whether the observed result matches the claim.

04

Concept explanation

Build a mechanism map that connects inputs, context, effects, and the reason they follow.

05

Product review

Pair visible product states with annotated findings, measured scores, and a verdict.

06

Skill evaluation

Connect the agent session, generated artifact, browser states, and evaluation evidence.

Explainer-native by design

A production contract agents can inspect

01ExplainerDocument

A structured source for captured and agent-authored explanations that compiles to editable TSX.

02Scene vocabulary

Whiteboards, diagrams, product walkthroughs, terminal sessions, browser evidence, and annotations.

03ExplanationBeat

One place for spoken phrases, visual actions, attention targets, media, and observed facts.

04Verification loop

Deterministic frame and timing checks that produce diagnostics an agent can act on.

Open source · MIT

Give your agent an explainer toolkit

Install the local renderer, add the Seqvio skill to your agent, and keep every explanation reviewable in the same repository as the work.

Renderernpm install -g @seqvio/renderer
Agent skillnpx skills add makesynt/seqvio
View Seqvio on GitHub