FiClaw data and model boundaries: desktop, cloud services and team pilots
Understand FiClaw Desktop, FiClawCloud, FiClawQuantAPI and the account portal before reviewing model context, data access and deployment requirements.
FiClaw content team · Published · Updated
阅读中文版Quick answer
FiClaw Desktop is the desktop workspace, FiClawCloud supports accounts, model routing and credits, and FiClawQuantAPI supplies data and quant interfaces. Installing a desktop app does not establish that every task runs offline. Confirm model context, external tool calls, data permissions and retention for the actual configuration.
For team leads, IT contacts and researchers evaluating a financial agent's data flow.
Separate the website, workspace and cloud services
The ficlaw.ai website provides product information, public downloads and inquiry forms. FiClaw Desktop is the installed research workspace. The FiClawWeb portal and FiClawCloud handle user-facing account pages and backend account, model and credit services respectively.
FiClawQuantAPI provides financial data and quant interfaces. Opening the desktop workspace does not mean every model or data call runs on the device. The actual path depends on configuration and the tools used by the task.
Review model context and data-tool calls separately
Document the context a model receives separately from the parameters sent to a data tool. Run a synthetic or redacted example through the pilot workflow and record file sources, tool calls, output locations and human-confirmation points.
Confirm data providers, permitted use, fields, model providers, processing regions and retention responsibility. Verify unresolved settings with the team contact; labels such as desktop or private deployment do not replace a review of actual data movement.
Evaluate public trials and institutional deployment separately
Public downloads support trying a clearly scoped task. Shared institutional agents, cross-role handoffs and organizational governance are development and pilot directions. Customer VPC, hybrid and private deployment require a separate assessment of data, permissions, interfaces and operations.
Agree on inputs, outputs, reviewers, permitted systems, exception handling and an exit plan. Authorized people retain responsibility for key conclusions, strategy deployment and fund-related operations. A roadmap is not a delivered institutional feature checklist.
Review responsibilities service by service
| Entry or service | Role | Check before the pilot |
|---|---|---|
| ficlaw.ai website | Product information, downloads and inquiries | Installer origin and inquiry form fields |
| FiClaw Desktop | Financial agent desktop workspace | Files, tool permissions and selected models |
| FiClawCloud / account portal | Accounts, model routing, credits and account pages | Model provider, context and recorded information |
| FiClawQuantAPI | Data and quant interfaces | Data permission, coverage, call parameters and output |
Frequently asked questions
Does FiClaw Desktop keep all task data on the device?
Do not infer this from the desktop format. Tasks may use cloud models, account services or quant interfaces. Check model context, tool parameters and transfers for the actual configuration.
Does the public installer provide a ready-made private deployment?
No. A public installer is not an institutional private deployment. VPC, hybrid and private setups need a separate assessment of data, permissions, interfaces, operations and acceptance criteria.
How is FiClaw related to SiStarAI?
FiClaw is operated and developed by SiStarAI, whose legal name is 硅星纪元（上海）人工智能科技有限公司. The product website is https://ficlaw.ai and the company website is https://sistarai.com.