An update shouldn't cost you your data
More and more people use AI to write standups and weekly updates. It saves time, but there's a question that's easy to skip: where do your screen, your chats and your drafts actually end up?
If you're a developer, product manager, designer or consultant, your screen is often full of unreleased plans, client files and internal discussions. A tool that syncs screenshots to the cloud by default is asking you to hand a full day of your digital footprint to a third party. That's not what you signed up for when you asked for help with an update.
Yesterly takes a clear position: a tool can understand your work without owning your data.
Cloud sync vs. local-first: what's the difference?
AI work trackers tend to fall into two groups:
- Cloud-synced: screenshots, notes and reports live on the vendor's servers. Switching devices is easy, but the vendor decides where the boundaries of your data are.
- Local-first: your work history stays on your device. The cloud only handles what it has to, like your account and plan.
Yesterly is local-first. Captures, notes, your timeline, recaps and their revisions are all saved on your computer. There's no cloud sync, and your work content is never stored on Yesterly's servers.
In practice, that means:
- Unreleased projects and client files don't get a new path to the cloud just because you wanted an update written;
- Browsing, editing and exporting all happen on your machine, and your history doesn't vanish because an account policy changed;
- You know exactly where your data goes, without decoding a long privacy policy.
Direct connections: requests skip our servers
Storing data locally is only half the story. The other half: how do AI requests leave your computer?
Yesterly connects directly to the AI you choose: the built-in model, your own API key, an AI subscription you already have (such as Codex, Claude or GitHub Copilot), or DeepSeek (web). When a capture needs a note or a recap needs writing, the request goes from your computer straight to that provider. Yesterly's servers don't proxy or relay it.
Our cloud only handles sign-in, device authorization, your plan and usage counts (such as tokens and number of images). It never receives your captures, notes or recap text. When you use a cloud model, that provider's own terms apply to what you send. Our privacy policy spells this out.
Or keep everything on your machine
If you'd rather your work never leave your device, point both the capture model and the recap model at a compatible model running on your computer. Recognition and recap writing then happen locally, with nothing sent to an outside AI service. (You'll need to run a local model with vision support yourself.)
It's an option for people with stricter privacy needs, not a default we push on everyone. The choice is yours.
Three questions to ask before you choose a tool
Before picking an AI work log, check:
- Are screenshots and notes stored on your device, or uploaded by default?
- Do AI requests go directly from the app, or through the vendor's servers?
- Can you review the sources before a report is written, and edit it afterward?
Yesterly's answer to the third: check the source notes before you create a recap, keep editing it afterward, and export it as Markdown or plain text. AI handles the tedious part. Judgment and wording stay with you.
The bottom line
AI can make status updates painless, but that doesn't mean handing over your entire working world. Local storage, direct connections and clear boundaries have guided Yesterly from day one. If you want a tool that helps you remember your work and spend less time writing about it, while respecting your data, download Yesterly and start with the free trial.