Field-level mapping and tracing
Connect leaf fields, then select one to inspect its sources, destinations, intermediate paths, and change impact.
About FieldWeft
See where every field comes from and where it goes
FieldWeft maps data flow between events, APIs, databases, and DTOs at field level. Record where a response value originated, how it changed along the way, and which domain or system boundary gives the surrounding nodes meaning. Add descriptions, tags, and typed metadata to preserve context that does not fit on the visible path itself.
The problem
A system diagram rarely tells you where a value came from
A single response often combines selected fields from several sources: a Kafka event, an external API, and a database query, for example. A service-and-table architecture diagram can show the overall system without answering where one particular value originated.
Mappings can live in documents or spreadsheets, but deep nesting and conditional inputs quickly make them hard to follow. FieldWeft closes that gap by representing entities as table-like nodes and field movement as edges.
Core capabilities
Focused tools for explaining data flow
Preserved and transformed values
Distinguish fields passed through unchanged from fields transformed along the way, with mapping styles and labels.
Processes and input variants
Represent transformations as black-box nodes and compare field structures that vary with a discriminator value.
Save, share, and back up
Autosave in the browser, then hand the graph to others through a share link or a backup file.
Your graphs are stored in the browser
Pages and graphs are stored in the current browser's IndexedDB. Lightweight preferences—the last-opened page reference, language choices, and canvas view settings—remain in localStorage. You can work without an account or a storage server, but clearing browser data can remove local copies, so export a backup file for important work.
A share link embeds the current page's complete FieldWeft document in the URL's #g=d1.… fragment. It preserves stable node, field, boundary, and relation IDs; explicit node relations and field mappings; descriptions, tags, and metadata; and exact positions, collapsed state, boundary membership, and sizes. The recipient explores it in a read-only view without saving, then chooses Import as a new page to save an independent copy. Wiki iframes open the import in a new app tab. Page names, other pages, and viewport state are not included, so use a backup file for several pages or very large graphs. A fragment is not sent to the web server or in a typical Referer header, but the full link can remain in browser history, clipboards, messages, and documents. FieldWeft v1 uses format: "fieldweft", version: 1, and five required arrays. Keep IDs stable when renaming or moving fields, and issue a new ID only for a new field. Never enter passwords, tokens, real customer data, or other sensitive values.
Who it is for
- Backend developers combining several APIs and events
- Engineers documenting field provenance in data pipelines
- Teams reviewing DTO impact and missing mappings together
- People tracing data movement through legacy systems
Current scope
FieldWeft does not analyze code automatically or monitor a live data pipeline. It is a lightweight documentation tool for drawing known structures directly or sharing validated FieldWeft v1 JSON. It does not provide server sync or collaborative editing, and it cannot reconstruct source SQL, real data values, or an external lineage repository. Share only the part of a graph needed for the review at hand.
Start with a sample graph
Explore it in read-only mode, then edit an independent copy.