Terms of Service

Last updated: April 2026

Acceptance

By accessing the geeViz Agent, you agree to these terms. If you do not agree, do not use the service.

Service description

The geeViz Agent is a research and analysis tool that provides conversational access to geospatial data via Google Earth Engine. A large language model (Google Gemini) writes and executes Python code on your behalf to produce maps, charts, thumbnails, and reports. The service is provided "as is" for research, education, and operational use by authorized users.

AI-generated content disclaimer

Analyses, maps, and reports are generated by AI using automated code execution against satellite imagery and geospatial datasets. Results may contain errors, use inappropriate methods, or misrepresent conditions. All outputs should be independently verified before use in any decision-making. This tool does not replace professional judgment.

Specifically, you should be aware that:

• Code may contain errors. The agent writes Python code based on its understanding of your request. This code may produce incorrect results, use wrong datasets, apply inappropriate methods, or miscalculate statistics. Always verify outputs independently.
• The model may hallucinate. The language model may generate plausible-sounding but incorrect explanations, cite nonexistent datasets, or produce fabricated statistics. Do not rely on the agent's text responses as authoritative.
• Results are not peer-reviewed. Generated analyses, maps, and reports have not been validated by subject-matter experts. They are starting points for investigation, not final products.
• Outputs may change. The same prompt may produce different results across sessions due to model updates, temperature settings, or changes in underlying datasets.

Acceptable use

• Use the agent for legitimate geospatial analysis, land management, environmental monitoring, and related research purposes.
• Do not attempt to bypass, disable, or circumvent the code sandbox, content filters, Model Armor screening, or any other security controls.
• Do not attempt prompt injection, jailbreak attacks, or social engineering to manipulate the agent's behavior.
• Do not use the service to access, exfiltrate, or probe for environment variables, API keys, credentials, or server configuration.
• Do not use the service to generate content that violates applicable laws or Google's Terms of Service.
• Do not redistribute or commercially resell outputs without appropriate attribution to the underlying data providers.
• Do not use the shared session feature to distribute harmful, misleading, or inappropriate content.

Security measures

The service implements multiple layers of security:

• Code sandbox — all user code and agent-generated code runs in a restricted environment. Filesystem access, network calls, subprocess execution, and dangerous builtins (eval, exec, open, os, sys) are blocked.
• Content screening — user messages and model responses are screened for prompt injection, code injection, data exfiltration, and social engineering via local regex filters and Google Cloud Model Armor.
• Parameter sanitization — custom script parameters use Python repr() for safe string escaping, preventing code injection via parameter values.
• Authentication — access is controlled via Google Identity-Aware Proxy (IAP). Only authorized Google accounts can use the service.

Despite these measures, no system is completely secure. If you discover a vulnerability, please report it to the developer rather than exploiting it.

Custom scripts

The custom scripts feature allows you to save and re-run parameterized code blocks. You should be aware that:

• Scripts are stored on the server and associated with your user account.
• Script parameter values are persisted and can be modified by anyone with access to your account.
• Scripts execute the same code sandbox as the chat agent — the same security restrictions apply.
• Script execution does not use the language model and incurs no LLM cost, but does consume Earth Engine compute quota.

Data and accuracy

The agent generates analyses based on satellite imagery and geospatial datasets. Results are provided for informational purposes only and should be independently verified before use in any decision-making process. The developers are not responsible for errors, omissions, or consequences arising from decisions made based on agent outputs.

Geospatial datasets have inherent limitations including spatial resolution, temporal coverage, classification accuracy, and cloud contamination. The agent does not warn about all such limitations.

Limitation of liability

The geeViz Agent is provided "as is" without warranties of any kind, express or implied, including but not limited to warranties of merchantability, fitness for a particular purpose, or accuracy of results. In no event shall the developers be liable for any direct, indirect, incidental, special, or consequential damages arising from the use of this service.

Availability

The service may be unavailable due to maintenance, updates, or infrastructure issues. We do not guarantee uptime, service continuity, or preservation of session data beyond the retention period.

Intellectual property

• geeViz — open-source software under the Apache 2.0 License.
• Your outputs — maps, charts, and analyses you generate are yours, subject to the terms of the underlying data providers (USGS, ESA, Google, etc.).
• Geospatial data — datasets accessed through the service are subject to the licensing terms of their respective providers. Many USGS and government datasets are public domain; others (e.g., commercial satellite imagery) may have restrictions.

Changes to these terms

We may update these terms at any time. Continued use of the service after changes constitutes acceptance of the revised terms. Material changes will be noted by updating the "Last updated" date at the top of this page.