Cluvra Cover Space

Cluvra

Not a tutorial. A living guide.

Cluvra is a desktop product idea that lets someone record a real workflow inside real software, package it into a guide, and let another person replay that workflow live on their own machine.

The core leap is simple: move beyond passive video tutorials into executable learning, where the learner can either study every step or walk away and come back to a finished file.

Pre-development Desktop Creator Economy Workflow Automation

Project Summary

Most software learning still happens through passive videos. You watch a workflow, try to remember it, rewind, miss steps, and rebuild it from memory inside your own app.

Cluvra proposes a different format: the living guide. A creator records a workflow once, and the learner can replay that guide directly inside their own software environment with controls for speed, chapters, annotations, and step-by-step study.

Two User Types, One Guide

User	What They Want	How They Use It
Output Seeker	The finished file with minimal effort.	Hits play, lets the guide run, and returns to a completed project file.
Learner	The technique and the finished file.	Watches carefully, uses slower playback or step mode, then studies and replicates the result.

How Cluvra Works

Record

The creator works normally in a target app while Cluvra captures the workflow through its bubble controls and overlay system.

Package

The recording becomes a guide file with metadata, chapter markers, annotations, and semantic labels.

Publish

The creator shares the guide as a free or paid link with profile attribution and guide metadata.

Replay

The learner runs the workflow inside their own copy of the software and chooses how actively to follow it.

Output

The learner ends up with a normal editable local file that belongs fully to them after playback finishes.

Floating Bubble

The floating bubble is the creator's recording control surface. It stays always on top, stays out of the way, and makes it possible to record without constantly returning to a main app window.

Recording State

A red pulsing dot signals active capture and expands into controls for pause, stop, chapters, and annotations.

Review Mode

A yellow state signals that Cluvra is still observing but holding actions for confirmation rather than silently dropping them.

Fast Control Surface

Elapsed time, action count, chapter insertion, and annotations all live in the bubble instead of the main window.

Review Mode Instead of True Pause

One of the strongest design decisions in the product is rejecting a naive pause button. If recording truly stops, the creator can make a critical change that the replay never sees, silently breaking the guide.

Silent Capture Buffer

While paused, Cluvra still records actions but marks them as unconfirmed.

Checklist Review Panel

On resume or stop, the creator quickly reviews a simple action list and chooses what belongs in the guide.

Human-readable Actions

Semantic labels turn raw clicks into understandable descriptions like selecting a layer or switching windows.

Guide Player

Automatic Replay

The guide executes on the learner's machine with no need for constant manual input.

Study Controls

Speed modes, chapter jumps, previews, and step mode support people who want to learn deeply.

Finished Local Output

The result is saved as a normal editable file on the learner's PC with no DRM layer attached.

AI Navigation Layer

Cluvra uses a three-tier replay strategy so recordings do not collapse the moment an app receives a minor UI update.

Tier	When It Runs	Behaviour
Exact Replay	Same app version	Runs raw coordinates directly for speed.
AI-assisted Replay	Minor version mismatch	Uses screenshots plus semantic intent to relocate the right UI elements.
Hard Stop	Major mismatch or missing element	Pauses and asks for manual help before continuing.

Creator Economy Angle

Free-to-paid Funnel

Creators use free guides to build audience, then convert that attention into premium guides, packs, or subscriptions.

Attribution in the Guide Layer

Credit lives in guide links, replay framing, annotation callouts, and remix chains rather than trying to lock the output file.

No-DRM Positioning

The output file is deliberately unrestricted, turning finished work into distribution and trust rather than friction.

Technical Direction

Layer	Choice	Role
Language	Python 3.11+	Fast prototyping for input capture, replay, and desktop logic.
Bubble	Tkinter first, Tauri later	Always-on-top recording controls with room to evolve visually.
Overlay Container	Tauri	Wraps the replay environment and control frame around the target app.
Replay	pyautogui + pynput	Executes guided actions during playback.
Vision Layer	Vision API fallback	Handles UI adaptation when software versions drift.
Storage	SQLite + Supabase	Supports local guide state, correction cache, profiles, and sharing.

Build Phases

Phase 0

Validate the core idea quickly: record clicks in one target app and replay them successfully.

Phase 1

Build the floating bubble, input capture, action model, review mode buffer, and basic replay engine.

Phase 2

Add checkpoints, version checks, AI-assisted adaptation, and broader cross-app testing.

Phase 3

Layer in speed modes, annotations, chapters, and learner progress features.

Phase 4+

Ship sharing, security, creator profiles, monetization, packaging, and launch loops.

Want to Talk About It?

Reach out if you want to discuss the format idea, desktop execution model, or the creator-economy logic behind Cluvra.

Get in Touch