Privacy Policy: Maestro Addendum
Last updated: October 2026
This addendum supplements the Orchestrel AI Privacy Policy for Maestro. It covers only what makes Maestro different. Everything else in the general Policy applies. If the two conflict, this addendum wins. Billing and cancellation specifics are in the Maestro terms addendum.
MP1. Local Execution
Maestro is a local orchestration platform: the AI agents and their tooling run on your machine, and the platform coordinates them over the network. Because the work happens where your code lives, the data Maestro processes is a small subset of what a hosted service would see. That difference drives everything in this addendum.
MP2. Data We Collect in Maestro
On top of the general categories in the Privacy Policy (account information, payment information), Maestro adds:
- Usage metrics: Dispatch counts, feature completion rates, build duration, and aggregate service usage statistics, used for billing and service improvement.
- Conversation and orchestration prompts: We store the prompts and responses exchanged with AI models so that conversations can be resumed and orchestration state preserved across sessions. Prompts are never used for training and are not shared with third parties beyond the AI providers required to fulfill your requests.
MP3. Data We Do NOT Collect
We want to be explicit about what we do not collect:
- Source code. All code execution, file generation, and code modification occurs on your local machine. No source code is uploaded to or stored on Orchestrel servers.
- File contents. We do not read, access, or store the contents of files in your development environment or project directories.
- Keystrokes or screen activity. We do not monitor, record, or capture your keyboard input, screen recording, or other local activity.
MP4. Prompt Retention
- Conversation and orchestration prompts are retained for the life of your active workspace, so sessions can be resumed, and are deleted within 30 days after the workspace or its sessions are removed.
- Account data, billing records, and usage metrics are retained per the Privacy Policy.
MP5. Third-Party Inference
Maestro accesses AI model providers via our routing infrastructure to run inference for your work. Prompts are encrypted in transit within the Maestro ecosystem, and we store conversation and orchestration prompts solely to ensure session continuity. We never sell, rent, or trade personal data, and we do not share it with third parties beyond the model providers required to run your requests. Model providers may process data on servers located outside the EU; we do not currently guarantee that all inference processing occurs within the EU.
MP6. Security Specifics
- No source code, file content, or local activity is ever transmitted to or stored on our servers
- Prompts are encrypted in transit within the Maestro ecosystem, in addition to TLS 1.3+ on all communications
General security measures (access controls, authentication, cookie policy) are described in the Privacy Policy.