WebCull
En
De
Fr
Es
Feature in planning

Link Metadata Progress

A public planning trail for turning bookmarks into structured, searchable data that people, tools, and AI agents can use.

Developing
12 public components
0% complete
How Link Metadata is taking shape 0%
Completed Planning 4 public components Selected Active Developing 4 public components Selected Upcoming Testing 4 public components Selected
Step 01 Completed

Planning

Planning covers the user problem, scope, constraints, success criteria, dependencies, and the ordered implementation plan established before development begins.

4 public components
WebCull
Bookmark organization
Completed
Component 01

Add structured information to each bookmark

Link Metadata will add custom fields and values alongside a bookmark's title, tags, and notes. A researcher could record an author and publication date, while a project team could track a status, source, owner, or reference number.

Creating a field on one bookmark will not create empty fields across the rest of a collection unless you want it to. Metadata should synchronize like other bookmark details, allowing the CLI to preserve structured information, internal state, collected facts, and other reusable data directly alongside a bookmark.

WebCull
Editing experience
Completed
Component 02

Make repeated fields quick to reuse

Metadata should remain flexible without making repeated organization tedious. When someone adds a field, WebCull can suggest field names they have used before, helping them reuse familiar labels without imposing a collection-wide schema.

The editing experience is planned as a compact part of Link Details, giving people greater control over what information appears on their bookmarks. Users will be able to hide or remove commonly unused details, such as descriptions or notes, across collections while preserving WebCull's core bookmark capabilities.

WebCull
Search and discovery
Completed
Component 03

Search bookmarks by field and value

Structured information becomes most useful when it can be searched directly. WebCull plans to let users find bookmarks that contain a particular field, match a value within that field, or are missing information they still need to collect.

The search experience may build on the syntax already used for tags, with new ways to target metadata keys and values. Exact and partial matches are part of the direction, while more advanced options such as regular-expression matching remain under consideration.

WebCull
Spreadsheets and datasets
Completed
Component 04

Use bookmark metadata in spreadsheets

Bookmark metadata can form reusable datasets that open naturally in spreadsheets. Users will be able to select bookmarks and fields, then export structured rows and columns for sorting, filtering, comparison, formulas, charts, and further analysis.

This can support research collections, inventories, project tracking, content reviews, competitive analysis, and other workflows where information gathered across many bookmarks needs to be studied or shared as a dataset.

Step 02 Active

Developing

Developing covers working product slices, implementation decisions, completed surfaces, open risks, and changes to the approved scope.

4 public components Current checkpoint 05
WebCull
Core implementation
Active
Component 05

Build the first complete metadata workflow

The first complete workflow will cover adding, editing, removing, loading, and synchronizing bookmark metadata while keeping changes current across sessions and devices.

It will establish the core behavior that search, spreadsheets, CLI access, reporting, richer value types, and other metadata capabilities build upon.

WebCull
Metadata management
Upcoming
Component 06

Manage metadata across all bookmarks or just one

WebCull will add a metadata manager for choosing what appears by default across all bookmarks. Users will be able to define reusable fields, hide information they do not need, restore it later, and remove custom metadata that is no longer useful.

The same controls will be available while editing an individual bookmark. A change can apply only to that bookmark or become an account-wide default, giving people control without forcing every bookmark into the same structure.

WebCull
Automation
Upcoming
Component 07

Give the CLI and AI agents full metadata access

The WebCull CLI is planned to support viewing, adding, changing, removing, and searching bookmark metadata, so the same structured information is available to scripts, terminal workflows, and other tools.

This is especially useful for AI agents and automation. Metadata can hold collected facts, processing state, identifiers, analysis results, and other internal information that scripts and agents may need to retain and revisit over time.

WebCull
Search and table view
Upcoming
Component 08

Search and organize metadata in a table

Search will gain a table view that turns matching bookmarks into focused rows and their metadata into columns. Users will be able to narrow the results with search, choose which fields to show, then sort and organize the information without leaving WebCull.

The table will also support direct metadata editing and spreadsheet-ready exports. Its first version will stay focused on bookmark research and organization rather than trying to reproduce every spreadsheet feature.

Step 03 Upcoming

Testing

Testing covers validation across supported environments, including defects, accessibility, privacy, packaging, copy, translation, and readiness for full launch.

4 public components
WebCull
Extensive dogfooding
Upcoming
Component 09

Dogfood Link Metadata on real product work

Dogfooding will be the first real test of Link Metadata, and this is one feature we are especially looking forward to using ourselves. One of the first projects will be organizing the demo videos used across the website, including where each video appears, what it shows, its production status, and where its source files live.

A deterministic CLI script will also keep the entire website synchronized with a bookmark collection as pages and links change. This will make the site searchable through the same tools used for any other collection, while putting metadata, search, and synchronization through a complete real-world workflow.

Link Metadata will also help organize the research, references, and other resources used while building the product. Testing it through everyday work should make gaps much easier to see and help shape what gets built next.

WebCull
End-to-end and unit tests
Upcoming
Component 10

Build full end-to-end browser and unit test coverage

Full end-to-end browser testing will cover the complete interface, from creating and managing fields to searching, editing table results, and exporting data. The same workflows will be tested across supported browsers, keyboard use, narrow layouts, and editable and read-only states.

Extensive unit testing will cover field validation, account defaults, bookmark-specific overrides, synchronization, conflict handling, CLI operations, and privacy boundaries. Together, this coverage should make future changes safer and catch broken connections earlier in development.

WebCull
Adversarial testing
Upcoming
Component 11

Challenge Link Metadata through adversarial testing

This is the phase where Link Metadata is treated as a target and actively attacked. Adversarial testing will probe every new storage, editing, search, table, CLI, and synchronization path for ways metadata could be exposed, altered without permission, or used to weaken an existing security boundary.

Link Metadata is designed to work for accounts that use end-to-end encryption and accounts that do not. Adversarial testing will cover both configurations, along with private, shared, and public collections, malformed inputs, authorization bypasses, cross-account access, stale clients, imports, exports, and attempts to turn metadata into a new attack surface. Findings from this work will guide security and privacy improvements as the feature continues to develop.

WebCull
Case studies and field learning
Upcoming
Component 12

Study Link Metadata in real-world workflows

Case studies will follow real work from metadata collection through search, organization, automation, and export. They will cover different needs such as research, inventories, project tracking, content reviews, and agent-assisted analysis.

Each study will show where Link Metadata is already useful, where the workflow still needs work, and what should be explored next as the feature continues to develop.