Begin with what you want to keep: a conversation, a photo library, years of notes or the records behind a personal project. You can build an archive a source at a time.
This workflow connects SEED's public guides with the collection experience planned for the ChatGPT plugin and vault. The guides are available now. Plugin and vault collection access follows the supported routes offered when access opens.
1. Decide what matters
Make a short list of the services, devices and existing files that hold the records you want. Note the account or profile, the period you care about, and whether original photos, recordings or attachments are essential.
You might start with:
| Your records | Example guides |
|---|---|
| Social posts and shared memories | Facebook, Instagram, TikTok |
| Messages and email | WhatsApp, Telegram, Gmail |
| Photos and files | Google Photos, iCloud Photos, Google Drive |
| Notes and AI conversations | Notion, ChatGPT, Claude |
| Stories and paper records | Record family stories, digitize paper records |
The directory is a starting point for a broad personal archive. Different providers expose different records, and some formats are easier to reuse than others.
2. Find the route for your source
Open the guide for the service. Check the account type, device, export options, available formats and limitations before requesting the files. Follow the provider's official page from the guide, especially if its menus have changed.
The ChatGPT plugin is designed to help identify that route from an ordinary question. A useful request is:
Help me preserve my Instagram posts and messages. Explain the export choices, link the official instructions and tell me what to check when the download arrives.
Guidance does not require access to the contents of your account. When a route depends on your device or account type, those details may be enough to choose the right instructions.
Learn about the ChatGPT plugin
3. Choose how to collect
Provider export. Request your copy through the provider and download it when it is ready. Some exports arrive in several parts or expire after a limited time. Keep every part and follow the source-specific guide.
Existing files. Gather archives, documents, photos and recordings you already control. Keep the original files and preserve their folder names and accompanying metadata where possible.
Supported connection or assisted collection. When SEED offers this route for your source, the collection flow identifies the account, requested information and destination before you approve it. Authentication belongs in the provider's or SEED's sign-in flow. Do not paste passwords, verification codes or provider access tokens into a chat.
A guide, an automatic download and a supported import are different capabilities. The guide remains useful when automation is unavailable. Check the route offered for the exact source instead of assuming every service in the directory can be collected automatically.
Compare collection methods · Plan storage and time
4. Check what arrived
Before you treat the collection as complete:
- Confirm that it belongs to the intended account or profile.
- Check that every archive part has downloaded and can be opened.
- Compare the received date range and categories with your request.
- Open a few representative messages, photos or recordings, including older ones.
- Look for missing attachments, unreadable files and links that depend on the original service.
- Keep a note of the export date, source, selected options and known gaps.
The provider may exclude information that was deleted, never retained or unavailable through that route. A successful download proves that you received a package; it does not prove that the package contains everything the account ever held.
5. Preserve the original and its context
Keep the original package as well as any copy you unpack for reading. Store the collection note beside it so you can understand the archive later. Maintain an independent copy of important records and check that you can open it.
When using a supported SEED import, upload through the authenticated SEED flow. The planned plugin journey uses a collection reference and a limited status receipt in chat; it does not require posting the raw archive into ChatGPT. A receipt should distinguish the files received from the material the service can currently organize or search.
Use the receiving service's supported route for sensitive records. A public preservation guide does not mean the contents of every archive are eligible to be sent through an AI plugin.
Understand data handling · Keep a copy you control
6. Choose what to use elsewhere
Preserving a collection does not automatically share it with another assistant. If you want help with a project, chapter or trip, choose the relevant context and the recipient separately.
A one-task brief can contain a fixed selection. An ongoing connection can make a chosen collection available for future questions under its stated permissions. Review that distinction before connecting a tool.
Use selected context through MCP · Build with the API · Understand permissions
If a source is missing or a step fails
Search the guide directory for the source and check its official references. If the provider has changed its route, you cannot open an archive part, or your account type is not covered, tell us the service and the step where you stopped. You do not need to send your archive or credentials to describe the problem.
Available guides and launch access
The public library currently contains 119 provider export guides and three direct-capture guides. It includes a wide range of personal records; it is not a promise that every provider, region, format or automated route is supported.
You can use the guides today without signup. Join the waitlist for the SEED collection, plugin and connected-vault experience.