Feedvault

Use case

Feedvault for directories and listings

If you maintain a list that others want to use (venues, suppliers, courses, events, resources), you probably share it as a spreadsheet that goes out of date the moment it is sent. An API with a key per partner keeps everyone on the current version and lets you see who uses it.

Reviewed 2026-11-16 · Feedvault team

Why not just share the spreadsheet?

  • Copies go stale: partners keep using last month's file.
  • You cannot see who uses it or how much.
  • You cannot withdraw access from one partner without changing it for everyone.
  • Partners have to write their own parsing for your columns.

An API with one key per partner fixes all four: everyone queries the current data, usage is metered per key, keys are revoked individually, and the response format is consistent JSON.

Data you are allowed to publish

Only publish data you own or are authorised to share. Feedvault imports files you upload; it does not scrape websites or enrich records with outside personal data. If your directory includes people (for example contact names), make sure you have their permission to share those fields, or leave them out of the dataset. Every dataset carries a provenance note so consumers know where the data comes from.

Set up the dataset

  1. Clean the sheet: one row per listing, a unique id column, consistent categories and country codes.
  2. Create a Feedvault dataset with a clear provenance note, upload the CSV or JSON and review the inferred schema.
  3. Mark the id as the key and the name, city and category columns as searchable.
  4. Create one API key per partner, each with a monthly quota that fits their expected use.

What partners can ask for

UseQuery
Venues in one cityfilter[city][ieq]=dubai
Several categoriesfilter[category][in]=cafe,bakery
Recently updatedfilter[updated][gte]=2026-11-01 & sort=-updated
Free-text searchq=rooftop

Responses are paginated with a cursor, so partners can page through large result sets reliably.

What a response looks like

Each response is JSON with the matching records, the dataset's name and provenance note, and a cursor for the next page when there are more results. Every record carries its key and the column values with their types: numbers as numbers, dates as ISO date strings, empty values as null. Partners can rely on the same field names on every request, because they come from the schema you reviewed.

Send partners a one-page note with the endpoint, the field list, three example queries and your contact. It saves a week of emails.

Communicating changes

  • Adding a column is safe for most consumers; renaming or removing one is not. Announce those changes in advance.
  • Keep the meaning of values stable: if "category" was free text, switching to codes is a breaking change.
  • Tell partners how often you update the data, and keep to it.
  • If you must make a breaking change, create a new dataset for the new format, give partners time to switch, then retire the old one and its keys.

Updates and change history

Re-import the file whenever the directory changes. Records are matched by the key, so updates replace the old values and new listings are added; Feedvault records the changes per dataset so you can see what moved between imports.

Governance

  • Write down what each partner may do with the data, and give each a separate key.
  • Set quotas so one heavy user cannot exhaust your plan.
  • Revoke a key when an agreement ends.
  • Keep the provenance note accurate: who maintains the data and how often it is updated.

Questions and answers

Can partners write to the dataset?

No. Partner keys query the data; you update it by importing new files.

Can I make the API public without keys?

No. Every query needs a key, which is what makes per-partner metering and revocation possible.

More Feedvault guides