Vinotify Logovinotify
Login

Automate Vinted Sourcing with Vinotify's API, MCP and Webhooks

← Back to Blog

8 September 2026

If you source Vinted items for resale, the useful automation is usually straightforward: keep your saved searches in Vinotify, read new matches into your own workflow, and decide what to inspect or buy yourself. Vinotify provides the managed search feed; your bot or coding agent can handle the rest of the hand-off.

The integration surface is available to Pro accounts. It includes account-scoped bearer tokens, a REST event feed, a direct MCP endpoint for compatible agents, outbound webhooks, and a durable event feed. Start with the full developer documentation when you are ready to connect a client.

Which one do you need?

There are three ways in, and they suit different jobs. Pick the row that matches what you are building; each links to the guide that covers it properly.

Choosing a Vinotify integration
If you want to…Use
Get matches pushed into a bot, spreadsheet or inventory tool the moment they appearWebhooks — the webhook guide
Ask an AI assistant or coding agent about your alerts in plain languageMCP — connecting MCP to AI tools
Run your own worker that can catch up after downtimeThe REST event feed — covered below and in the developer documentation

A reseller workflow that stays understandable

  1. Create saved searches for the items, brands, sizes, prices, and keywords you actually source.
  2. Create a read-only integration token in Integrations and keep it in your local environment or secret store.
  3. Discover active searches, then read each event stream from a saved cursor.
  4. Deduplicate by event_id, turn a match into your own inventory or review record, and open the Vinted URL when you want to inspect it.

Event delivery is at least once. That is useful for a resilient bot, but it means your consumer should save next_cursor and ignore a repeated event_id it has already handled.

REST for a small poller or bot

A coding agent can build a small worker around two calls: refresh active searches with /api/v1/integrations/discovery/v2, then long-poll /api/v1/integrations/events?search_id=…&cursor=…&wait=30. The response gives you the events and the next cursor for that search.

Authorization: Bearer $VINOTIFY_TOKEN

GET https://vinotify.me/api/v1/integrations/discovery/v2
GET https://vinotify.me/api/v1/integrations/events?search_id=42&cursor=0-0&wait=30

Keep the cursor per search. If the feed says the cursor has expired, reset to the returned reset_cursor value. If the API returns 429, wait for Retry-After; if it returns 503, back off without discarding the saved cursor.

MCP for a coding agent, in short

MCP is a good fit when you want an agent to inspect your alerts in natural language. Point a compatible client at https://vinotify.me/api/v1/mcp and add the bearer token in that client's local configuration. The read tools are list_searches, get_new_items, and whats_new.

Write tools are separate and narrow. A token can be created with write access, and the server must also have its write feature enabled, before an agent can create, update, pause, or resume a search. There is no delete tool. Leave the token read-only unless you have a clear reason to delegate search changes.

{
  "mcpServers": {
    "vinotify": {
      "url": "https://vinotify.me/api/v1/mcp",
      "headers": {
        "Authorization": "Bearer YOUR_TOKEN"
      }
    }
  }
}

Client configuration, the full tool list and the token model are covered in connecting Vinotify's MCP server to AI tools.

Push or pull, in short

Push is the fastest route. Set a destination on a Pro search with PUT /api/v1/searches/<search_id>/webhook/external-url and Vinotify POSTs a new_items body to it the moment a search finds something. The destination has to be a public HTTPS endpoint, and Vinotify checks it both when you save it and again before every delivery.

Push is best effort, though. There is no retry queue, so if your receiver is down the delivery is gone. The payload stays available for 24 hours at /api/v1/searches/<search_id>/webhook/latest, and the bearer-authenticated event feed under /api/v1/integrations keeps 48 hours of history behind a cursor. Run the feed alongside push if you cannot afford to miss an item, and deduplicate on event_id.

Payload shape, retry behaviour and worked setups are covered in the webhook automation guide.

Keep the boundary clear

Vinotify supplies listing evidence such as the captured time, asking price and currency, title, URL, item id, and available item details. It does not buy items, require Vinted credentials, or turn an alert into a confirmed sale. Your own workflow can score or queue candidates, but keep those decisions separate from what the event actually proves.

Store the token outside source control, refresh discovery as your searches change, and revoke the token from Integrations when you no longer need it. For exact parameters, limits, response fields, and errors, use the Vinotify developer docs.

Frequently asked questions

Does Vinotify have an API?
Yes. Pro accounts can create account-scoped bearer tokens and read new Vinted matches through a REST event feed at /api/v1/integrations, an MCP endpoint at /api/v1/mcp, or outbound webhooks that POST to a URL you own. Full parameters and limits are in the developer documentation at vinotify.me/docs.
Do I have to give Vinotify my Vinted login?
No. Vinotify never asks for Vinted credentials or cookies and does not sign in to your Vinted account. It watches public listings for your saved searches and reports matches, so the integration surface carries no risk to your Vinted account.
How quickly does a bot hear about a new listing?
Paid searches are checked every minute, and a matching item is typically discovered within 0.5 to 0.6 minutes. Push webhooks deliver as soon as the match is found; the event feed supports a long poll of up to 30 seconds so a consumer can wait rather than poll in a tight loop.
Can my bot create and manage searches, or only read them?
Tokens are read-only by default. Write access is a separate scope chosen when the token is created, and the write tools can create, update, pause, and resume searches. Writes are budgeted at 10 a minute and 100 a day per token.
What are the rate limits?
240 requests a minute per token, at most 12 concurrent long polls per token, and for write-scoped tokens 10 writes a minute and 100 writes a day. Exceeding a limit returns 429 with a Retry-After header.
Which plan do I need?
The integration surface — tokens, REST, MCP and webhooks — is available on the Pro plan.