Legal
Subprocessors
Last updated: October 26, 2026
These third parties process customer data to provide SynthPlanner. The chemical structures and text we send to the AI model providers and chemistry services do not include your name, email address, organization name or account ID. We give customers at least 10 business days' notice before adding a subprocessor, and customers may object within 30 days.
| Provider | Purpose | Data it receives | Location |
|---|---|---|---|
| Google Cloud | Hosts the service: servers, databases, cache, file storage, logs and secrets | All data in the service | United States (us-central1; some file storage and backups in Google's US multi-region) |
| Anthropic | Claude models: checks and ranks routes, proposes alternatives, fills in reaction conditions, estimates quote labor, reasons about supplier lead times, runs the planning assistant | Molecule and reaction structures, conditions, reagents, procedure text, your briefs, chat messages and the assistant's fetched results | United States |
| Google Vertex AI | Gemini models: suggests lab equipment for a reaction step | Reaction and product structures, condition text with reagent names | United States (us-central1) |
| Chemspace | Supplier availability and pricing | Molecule structure and ship-to country | — |
| Molport | Supplier availability and pricing | Molecule structure | — |
| CAS Common Chemistry | Compound identification | Molecule structure and CAS registry number | — |
| PubChem (US National Library of Medicine) | Compound names, identifiers and properties | Molecule structures, identifiers and names | United States |
| OPSIN (EMBL-EBI) | Checks chemical names | Chemical names | — |
Model providers' terms bar them from training on our data. Anthropic deletes API inputs and outputs within 30 days; content its safety systems flag can be kept up to 2 years. Google keeps a prompt only if its abuse systems flag it, for up to 90 days. We do not have zero-retention agreements with these providers.
Not used by the production service today: OpenAI and Mistral AI (pre-release notebook search and document reading), and PubMed, Europe PMC, Crossref and OpenAlex (pre-release literature search). We will update this page before any of them receives data from the production service.
Our website and business tools (Base44, Google Fonts, unpkg.com, Calendly, Google Workspace, Notion, Origami and Clean) receive website and business contact data, not customer scientific data.