Save more, earn more
Mac compute first. Estimated savings of 40%–70%.
GouziYour local AI creative workspace on Mac
Let Codex do the work. You approve.
Local first. Cloud on demand. Connect Codex via MCP to create with text, images, video and audio in one place.Free for individuals. Teams share APIs and local compute.
Requires macOS 27+
Build lasting assets
Data is valuable in the AI era. Keep it local when possible.
Your own intelligence
Use accumulated private data to train dedicated AI.
AI is being used. The value is leaking away.
Teams use AI every day, but accounts, keys, projects, assets, and methods remain scattered across tools and individuals. The team keeps investing without building creative assets it can reuse and carry forward.
Scattered accounts
Tools, settings, assets, and results stay disconnected across people.
Exposed keys
Keys copied to member devices are hard to rotate, revoke, and hand over.
Duplicate spend
Cloud services are still called for work the Mac can already do.
Lost know-how
Useful choices disappear with personal accounts, tools, or team changes.
Tools can change at any time. The team should not have to start from zero every time.
The critical choices stay with your team.
Gouzi does not choose the model for you. Your team decides how compute is used, which models to connect, where assets stay, and how capability is shared.
Your compute
Use system capabilities, local models, and on-device media processing first. Call cloud models only when a task needs them.
Your models
Connect the team’s own model APIs. Models can change without forcing projects and canvases to migrate.
Your assets
Projects, source material, versions, and AI results stay on team devices by default for continued editing and reuse.
Your team
Members use shared team AI capabilities without access to team keys or private configuration.
03 / HOW LOCAL AND CLOUD DIVIDE THE WORK
Local is the home ground.
Local-first does not mean entirely offline. Your team decides when to use on-device capabilities and when to call the cloud models it selected.
Finish locally first
Editing, layout, media processing, and local models
Use cloud when needed
Send only selected specialist generation tasks to cloud models
The team gets access. The keys stay put.
An administrator sets up shared AI capabilities so everyone can create without sharing API keys or private settings.
Admin managed
Shared AI capability
Safe team accessKeys stay privateTeam setup
Manage the AI capabilities available to the team.
Keys distributed
Members use capability without handling team keys.
Access control
The team controls member and device access.
iAn administrator manages shared capabilities and adjusts member and device access as needed.
Real capability, visible.
These capabilities are available now, keeping text, images, video, and audio editable and reusable in the same project.
OBJECT AI4 object types
Select an object. Work on it directly.
Text, images, video, and audio show only relevant tasks. Results return to the canvas to compare, accept, reject, or undo.
GOUZI VIDEO8 reviewable steps
From product material to marketing video
From reference and product analysis through script, storyboard, sound, generation, review, and assembly, every step can be checked.
PRO SKILLS3 skill groups
Handle specialist tasks in one workspace
TypeCraft, Xiaohongshu covers, video translation, and smart composition share one local project, task system, and model boundary.
Models change. Projects and creative assets remain with the team.
Clear data boundaries
Three boundaries matter: where projects stay, who controls credentials, and when cloud services are called.
01
Content
Projects and creative assets
Default locationLocal devices→Stored locally by default
02
Content
Model credentials
Default locationAdmin managed→Not exposed to team members
03
Content
Explicit cloud tasks
Default locationThe chosen model provider→Only task-required content is transferred
Gouzi makes the boundary visible. Your team still decides whether to use cloud services and which provider to call.
Straight answers to important questions
01Does Gouzi lock us to one model?＋
No. Your team can combine local capabilities with cloud models it chooses.
02What works without a network?＋
Local editing, saving, export, and installed local capabilities. Cloud tasks and team-shared cloud APIs require a network.
03Are projects uploaded by default?＋
No. Projects and assets remain local by default. Explicit cloud tasks transfer only what the current task needs.
04Can members see team keys?＋
No. Members use approved AI capabilities without access to keys or private configuration.
05How are personal and team plans priced?＋
All personal creation features are free and local-first. Your first team starts a 30-day, 3-seat trial including the administrator. Expiry blocks new team operations, not existing data. Activate or renew with an activation code. Configure your own cloud APIs and pay providers directly. Gouzi does not currently resell cloud credits or offer online payment. For team services, contact us on WeChat: gouzi-me.
GOUZI FOR TEAMS
A local-first AI creative workspace
Requires macOS 27 or later. Free for individuals; teams get a one-month free trial and share APIs and local compute.
Download free macOS App