# Changelog of Instagram Scraper: Profiles, Posts, Reels & People Search (`stefanie-rink/instagram-intelligence-mvp`) Actor

- **URL**: https://apify.com/stefanie-rink/instagram-intelligence-mvp/changelog.md
- **Full Actor documentation**: https://apify.com/stefanie-rink/instagram-intelligence-mvp.md

## Changelog

### October 22, 2026

- Fix a feed error that stopped collection after the profile record.
- Include required query flags whose names do not end in `Enabled`.
- Add the required `PolarisSEOIsCrawler` flag to the fallback feed queries.
- Verify the first feed page and pagination with regression tests and live Instagram requests.

### October 19, 2026

- Add optional error reporting with original source locations and safe operation indexes.
- Add account pseudonyms and bounded input summaries without queries, credentials, or profile names.
- Preserve request delays, cooldowns, proxy rotation, result limits, and billing.

### 0.3.1 — October 16, 2026

- Replace the retired people-search endpoint with the current authenticated web GraphQL query.
- Discover query IDs and required feature flags from the current Instagram web scripts at startup.
- Update the fallback queries for profiles and feed pages.
- Request full feed pages to avoid empty responses for small result limits.
- Keep session cookies and CSRF headers out of requests to the public script CDN.
- Add regression tests for managed search, query discovery, feed pagination, and malformed search responses.

### 0.3.0 — October 10, 2026

- Include a managed Instagram session and proxy pool in the published Actor.
- Keep managed cookies out of customer stores and customer-controlled proxy connections.
- Prefer customer cookies and saved sessions when supplied.
- Stop network work when the result spending budget is exhausted.
- Count and remember content only after the SDK saves a result.
- Apply the configured timeout directly to the native HTTP transport.
- Add release checks for credential isolation and the pinned SDK's automatic result billing.

### 0.2.0 — October 10, 2026

- Download the Webshare proxy list and refresh it during runs.
- Rotate blocked or unavailable proxies with bounded attempts and per-proxy cooldowns.
- Preserve session cookies, the user agent, and the preferred proxy across runs.
- Use Chrome TLS and HTTP/2 profiles with certificate validation.
- Replace obsolete profile and feed routes with current web GraphQL queries and fresh web tokens.
- Add feed pagination, carousel URLs, current media types, and nullable source counts.
- Validate input before network work and verify returned profile identity.
- Give explicit profiles priority and preserve search metadata on profile and content records.
- Apply a total result limit and preserve correct state for only-new filtering.
- Keep diagnostics and search outcomes outside the result dataset.
- Fail blocked runs and storage errors while preserving earlier output.
- Define valid Actor, input, output, and dataset schemas.
- Exclude local credentials and storage from uploads and Docker images.