Build a shared AI workspace, turn repeatable tasks into skills, and develop agents or applications with China’s Coze service.
Coze, known as 扣子 in China, combines an AI work environment with tools for building agents and applications. This profile focuses on coze.cn. Its documentation describes projects where people, agents and files can stay together across ongoing work. Coze Programming is a related development environment for turning requirements into software and workflows. Choose the workspace when the deliverable is work; choose the development tools when you need a reusable system.
International readers should use the separate coze.com entry and check the service available to their account. This review does not establish identical subscriptions, shared accounts or feature parity between the two domains. A domestic plan’s model list or credit allowance should not be applied to the international service.
Prepare a weekly operations brief
Create a reusable research workflow
Prototype a small internal application
Coordinate a content-production project
Official product image. Click to inspect the details.
These capabilities span the workspace and programming products. The official guides distinguish their interfaces; do not assume every feature is available in every conversation mode.
Define one deliverable and its acceptance rules
Create a project with a small source set
Review an outline and a working draft
Test the result, permissions and credit use
Sample prompt: “Using only these meeting notes and approved metrics, draft a one-page weekly operations brief for our team lead. Separate results, risks and decisions. Cite the source file for every number. Mark missing facts as unknown. Show an outline first; do not send messages or change external records.”
The Chinese service offers free access and paid personal, team and enterprise editions. Entitlements can cover usage, collaboration, models and deployment resources.
The app FAQ explains that tasks consume credits for underlying resources such as models, compute and third-party services. Therefore, one request is not a fixed-cost unit. Compare a representative task including revisions. Budget separately for your own model credentials or external services, and do not assume a membership includes unlimited production traffic.
The listing’s main link and pricing discussion concern coze.cn. For coze.com, verify sign-in, regional availability and entitlements there; do not transfer domestic plan assumptions.
Workspace tasks can begin with a brief. Building an application benefits from understanding data, permissions and failure cases, even when the system generates the code.
Supported MCP, plugin, API and SDK routes provide integration options. Start with a narrowly scoped test connection and confirm what the connector can read or change.
No. Free access has limits, while paid editions and resource consumption introduce charges. Inspect the account’s usage screen after a small realistic task.
No. Check factual claims, spreadsheet formulas, presentation layout and application behavior in their actual output format. Keep the original inputs available for comparison.
No. Coze Studio is a separately deployed software project. Hosting, model calls, operations and commercial service subscriptions are distinct costs.
Confirm permission, sharing scope and training preferences. Remove unnecessary personal information and review the policies of any connected third-party service.
Explore platforms, inputs and outputs, licensing, and access requirements.
These are FindGoodAI workflow suggestions, not promised results or measured productivity gains. Start with work you can independently evaluate.
Official documentation covers web, Windows/macOS desktop and mobile access. Account registration is needed for the complete experience; the domestic registration guide lists phone and supported Chinese identity-provider routes. Desktop tasks involving local files depend on device access and granted permissions.
Inputs may include instructions, project files and authorized connected data. Outputs vary by mode: written material, work artifacts, executable workflows or applications. Confirm the actual download format and deployment target before preparing a large job. The development guide describes browser-based coding and API integration, but a generated app is not proof that its business logic is correct.
Coze Studio is a separate Apache-2.0-licensed repository for building and running agents and workflows. It is not a license to self-host the entire commercial website. Running it requires model configuration, infrastructure and maintenance; third-party models retain their own terms and costs.
The personal subscription agreement describes an account setting controlling use of data to improve the experience. Review that setting and the agreement applicable to your account before uploading sensitive work. A plan label alone does not establish that every connected provider excludes training.
The service terms and developer data-processing addendum distinguish platform obligations from your responsibilities for uploaded data and third-party connections. Keep source-material rights, generated-content review and permission to publish separate. For local tasks, select a limited folder and retain originals; for integrations, keep credentials outside prompts.
Checked on October 3, 2026.
App, credits and account FAQ ↗
Personal subscription agreement ↗
Loading experiences…
Saved tools are private. Approved comments are public; edits return to moderation. Comment counts include approved comments and replies. Each person has one rating. New ratings are automatically approved; withdrawn, unapproved or invalid ratings do not count. Share real experiences and avoid spam, private information or personal attacks.
For moderation appeals or data requests: support@findgoodai.com
Bring a real task and see how it fits the way you work.
User experiences