One collector, two surfaces, one evidence model

The Collector: collecting in your own browser

No Instagram login of ours serves your jobs, and no third-party collection provider is involved. Every post is collected by the Power Scraping Collector browser extension, inside your own browser, under your own Instagram login, from your own device and network. Your Instagram password and session never leave your browser: we never receive, store or replay them. The extension has no cookies permission and never reads cookies. It has no standing access to instagram.com either: in manual mode it reads only the tab you click it on, and automatic mode needs a permission to open instagram.com, which the Collector asks for when you allow automatic collection and gives up whenever automatic collection is off.

1. Install

Use desktop Chrome, Edge or Brave. During the pilot the Collector is installed from a download rather than from a store:

1. Download power-scraping-collector.zip (also linked from the dashboard's Collector panel) and unzip it into a folder you will keep.
2. Open the browser's extensions page: chrome://extensions, edge://extensions or brave://extensions.
3. Turn on Developer mode.
4. Choose Load unpacked and select the unzipped folder.

A Chrome Web Store listing is coming. Phones, Safari and Firefox cannot collect; any browser can still read, search and download results.

2. Pair it with your account

In the dashboard's Collector panel, create a pairing code. It is short, single-use and expires after 10 minutes. Paste it into the Collector (its icon in the browser toolbar), check that it names your organisation, and confirm. The paired Collector works for the person who created the code: it collects the jobs that person submits, including jobs submitted with an API key that person created.

To stop using a browser, press Disconnect this browser in the Collector or Disconnect in the dashboard; either revokes it at once. We delete a disconnected browser's record - the name you gave it, the Instagram handle it reported and its timestamps - 30 days after it is disconnected, and the Instagram handle recorded against each job 30 days after the job finishes.

3. Collect: manual or automatic

Either way, every job needs your click to start. A job waiting for your browser shows Waiting for your browser in the dashboard until you open the Collector and press Start. Once started, a manual job stays with your browser for 4 hours, renewed while the Collector is in use, and an automatic job for 30 minutes. The Collector stops at the first Instagram warning, challenge or logout. A challenge pauses that browser for 24 hours and switches automatic mode off until you turn it back on.

For a target declared as your own account (own_account), the Collector collects only while the browser is signed in to Instagram as that account - for Send this post as well as for jobs. A post already sent in the last 24 hours is a duplicate: it is not added again and not metered again.

What it sends, and what it never sends

Only these fields of the declared target's posts: post id, handle, permalink, dates, caption, location label, media type, links to the media on Instagram's servers, and like and comment counts. It never sends comments, other page text, your feed, your messages or who you follow. Posts of any other account are dropped on your device and again on our server. We never download the photos or videos themselves: records keep the links only. We then enrich, meter, store and export those records under your plan's retention rules.

Your account, your responsibility. Automated activity may trigger Instagram warnings, temporary action blocks or restrictions on your Instagram account. You must follow Instagram's terms, and you must have a lawful basis for every profile you collect - one of the four declared authorization bases below.

Step 0: declare what you may collect

This is not paperwork you can defer. Collection is refused with collection_not_authorized until two things are true: your account has accepted the current acceptable-use version, and the handle is an active authorized target with a basis and a reference. The Collector also drops, on your device and again on our server, any post that belongs to another account.

Revoke a target the moment its basis stops being true. Revocation takes effect immediately: the next job for that handle is refused. If the owner of a profile objects to its collection, we revoke it for them: that profile is then suppressed across the whole service and cannot be declared again, by you or by any other customer.

Reading a record and its citations

Each record carries its evidence items and the citations that bind fields to them. A field with no resolvable citation is not written at all, so a null means "we could not evidence this", never "we did not get round to it".

{
  "record_id": "rec_8d31f0a4",
  "source_url": "https://www.instagram.com/p/Cx1y2z3/",
  "creator_handle": "trattoriadaoscar",
  "venue": {"name": "Trattoria da Oscar", "place_id": null},
  "location": {"city": "Milan", "confidence": 0.94},
  "assessment": {"sentiment": "positive", "explicit_rating": null},
  "price": {"mentions_eur": [14.0, 22.0], "signal": "mid_range"},
  "unknown_fields": ["assessment.explicit_rating"],
  "citations": [
    {
      "field": "venue.name",
      "evidence_id": "ev_01",
      "modality": "caption",
      "excerpt": "finally back at @trattoriadaoscar for the cotoletta",
      "method": "caption_mention",
      "confidence": 0.97
    },
    {
      "field": "price.mentions_eur",
      "evidence_id": "ev_03",
      "modality": "caption",
      "excerpt": "cotoletta 22 euro, calice di rosso 14",
      "method": "price_extraction",
      "confidence": 0.93
    }
  ]
}

Every job also records what was collected, when, and under which declared authorization basis, and every field carries the evidence it was derived from. Download every record of a job, citations included, from GET /v1/jobs/{job_id}/export?format=csv or ?format=jsonl.

MCP connector quickstart

The connector exposes the same evidence-grounded records to an AI client over the Model Context Protocol, Streamable HTTP transport, protocol revision 2025-06-18.

Collecting from a chat needs the Collector. Before collection works, the chat user needs the Power Scraping Collector installed and paired on a desktop browser (Chrome, Edge or Brave). Until then start_collection answers collector_required, with a link to set it up. Reading and searching records works from any client, phones included.

Add the server

In Claude, ChatGPT or an IDE agent that supports remote MCP servers, add a connector with this URL:

https://api.powerscraping.app/mcp

The client discovers authorization by itself: the server answers an unauthenticated call with 401 and a WWW-Authenticate header pointing at its RFC 9728 metadata. Registration, consent and token exchange then happen without you pasting a secret anywhere.

The tools

A typical conversation

1. list_authorized_targets to see what may be collected.
2. authorize_target if the handle is missing — the basis is required.
3. get_usage before a large run, so the job is not refused mid-way.
4. start_collection, then get_job_status until it succeeds. If it answers collector_required, follow the link, pair the Collector on a desktop browser, and ask again.
5. get_venue_records or search_records to answer the question.

Tell the assistant to quote citations rather than fill gaps. When a field comes back null it is genuinely unknown, and the whole point of this connector is that the model can say so.