SieveHow Sieve protects your lists
You're trusting us with your contacts. Here is exactly how we keep them private, and the controls you have over them.
Every email address you upload, and its normalized form, is encrypted with AES-256-GCM before it is stored. When you upload a spreadsheet, the rest of your file is encrypted with the same cipher. The plaintext addresses are never written to disk in our database.
Your jobs, results and stats are scoped to your account. There is no shared view: you can only ever read and manage data that belongs to you.
Delete any individual job, its results and the file you uploaded from its detail page, or wipe every job tied to your account at once from the Account page. Deletion is immediate and permanent.
The validation pipeline processes addresses in memory. We never write raw email addresses into application logs.
For each validation job we keep the encrypted addresses, their per-layer validation verdicts and a confidence score, plus job metadata such as the file name and timestamps. This lets you revisit results and export them later.
When you upload a spreadsheet we also keep the file itself, all of its columns, encrypted with the same cipher as the addresses. We keep it for one reason: so your export can hand your own file back to you with a single status column added, instead of a reduced list of addresses. Those columns are never used for anything else. Your file is deleted when you delete the job, when you wipe your account data, and automatically at the end of the retention window below.
Validation requires querying public DNS and, where available, mail servers for the address's domain. We never sell your lists or pool them across customers. Addresses leave Sieve for a third party in only one case: the optional Pro AI tools you choose to run (the contact finder and the outreach drafter) send the sampled addresses to the AI and search providers named below so they can do their work. Plain validation never does this.
Some processing relies on trusted providers. Where a provider sits outside your region, running that feature transfers the listed data across borders.
- Paddle (Merchant of Record, UK and EU) handles all checkout and card data. Card numbers never reach Sieve; we keep only billing identifiers (your Paddle customer and subscription ids) and your credit balance.
- Anthropic (United States) receives the sampled addresses only when you run the AI contact finder or outreach drafter.
- SerpAPI and Google (United States) receive a sampled address or name as a search query only when you run the AI contact finder.
- Gravatar (Automattic, United States) receives a one way hash of an address, never the address itself, during optional web presence checks.
- Hosting and email infrastructure process data in the deployment region to run validation and send transactional email.
Plain email validation never uses Anthropic or SerpAPI. The operator can disable the AI tools entirely for privacy sensitive deployments.
You are the controller of the contacts you upload and Sieve is your processor for them; for your own account details Sieve is the controller. You can access, export and erase your data at any time.
- Access and portability: download a full copy of your data as JSON from the Account page.
- Erasure: delete any job, wipe every list at once, or close your account entirely. Deletion is immediate and permanent.
- Retention: uploaded lists, and the spreadsheet they came from, are deleted automatically after the retention window configured for this deployment. On sievemails.com that window is 90 days. The job itself stays in your history with its name, counts and dates, but once that window passes there is nothing left to export. Billing records are kept only as long as financial record law requires.
For any data request or question, contact [email protected].