Floe Agent
Docs

Disclaimer and usage notice

For Floe Agent, its Bridge companion apps and the project website. Updated: 4 November 2026.

1. The project and model services

Floe is an individually maintained open-source client and workspace for on-device files, editing, execution environments and connections to users’ own devices. The project website provides product information, documentation, download guidance and voluntarily submitted feedback; it does not provide online AI inference.

For cloud AI, users choose, obtain authorization for and configure their own provider, account, API key, endpoint and model. The client connects according to that configuration. The maintainer does not provide a shared AI model API, public model hosting, account top-ups or shared paid accounts. Users may also choose or download supported models to run on-device, subject to each model’s license and conditions.

Provider presets, compatibility lists and examples do not establish a partnership, agency or endorsement. Third-party prices, quotas, availability, terms and data practices are governed by the relevant provider.

2. AI output and verification

Generated text, images, audio, video, code and other outputs may contain errors, omissions, bias, fabricated information or disputed rights. Outputs do not express the maintainer’s views, promises or professional advice. The maintainer does not review each input or output of models configured by users.

Before using, executing, publishing, distributing or commercializing output, users should verify its accuracy, lawfulness, permissions and intellectual-property status. Do not rely solely on AI output for medical, legal, financial or other professional decisions. Apply legally required generated-content labels where relevant.

The responsibilities of providers, users and other parties depend on their actual conduct, service relationships and applicable law. The fact that content was generated by AI does not, by itself, place all responsibility on one party or release every party from responsibility.

3. Data, credentials and third-party processing

Cloud models, speech, search and other third-party services may receive task inputs, selected files and relevant tool results. An on-device workspace does not make every feature offline or establish a third-party promise of zero data retention.

Ensure you have permission to process, upload and share material. Review provider terms and access scopes, protect accounts and API keys, and do not submit unauthorized personal information, trade secrets or other people’s credentials. Applications, feedback and diagnostics voluntarily submitted to the project are used for the corresponding contact and support purposes; redact them before submission. This notice does not replace applicable project or provider privacy information.

4. Tool execution and lawful use

Use the software only with files, devices, networks and accounts you are authorized to access. File changes, code, terminal commands and remote connections can have real effects. Check task scope and results, and back up important material.

Do not use the project for unlawful activity, fraud, infringement or bypassing another party’s authorization and security boundaries. A software license, filing or this notice is not a certification that a particular use or generated result is lawful, accurate or safe.

5. Licenses and limits of responsibility

Software and related material are provided under their applicable open-source licenses. Third-party components and models retain their own licenses. This notice does not change those licenses, grant additional rights to third-party content or replace provider terms.

To the extent permitted by applicable law and licenses, the maintainer gives no additional guarantees for third-party model services or generated outputs. Nothing here excludes or limits liability that cannot lawfully be excluded or limited, or restricts users’ statutory rights. The project’s classification, duties and responsibilities depend on actual operations, applicable law and decisions of competent authorities; this page does not replace legal obligations.

6. Contact and feedback

Use the channels in Support for software issues, rights concerns or security risks. Do not post API keys, passwords or unredacted personal information in public reports.

Related reference: China’s Interim Measures for the Management of Generative AI Services. See also the Civil Code of the People’s Republic of China, including its statutory limits on exclusions of liability. Applicability depends on current rules and the actual circumstances.

简体中文

Search documentation

    Results open the matching section.