For app-only operations (search, read-only analytics), use the Bearer token directly — simpler authentication but no user-context actions.

Phase 2

Build the posting pipeline. The v2 POST /2/tweets endpoint accepts text, media IDs, reply settings, geo, and poll options. Media uploads use the v1.1 media/upload endpoint (still required in 2026 — v2 media upload is not yet available) and return media IDs that you attach to the tweet. Long-form posts require the note_tweet parameter with the full text — the 280-character limit applies only to legacy clients.

Implement rate limit handling at the endpoint level. Each v2 endpoint has its own rate limit (typically 300 requests per 15 minutes per user for posting, 450 per 15 minutes for app-only reads). Use a token bucket algorithm with per-endpoint tracking. Nakrut.pro's internal scheduler tracks remaining quota per account and per endpoint, queueing posts when limits are hit.

Schedule posts using the API's native scheduling capability (POST /2/tweets with scheduled_at parameter). This stores the post on Twitter's servers and posts at the scheduled time, eliminating the need for client-side cron jobs. Note: scheduled posts do not count against rate limits until they're actually published.

Phase 3

Set up the filtered stream for real-time monitoring. The v2 POST /2/tweets/search/stream endpoint lets you define rules (e.g., "@brandname OR (brandname context:55.*)") and receive matching tweets in real time via a persistent connection. Add rules via the POST /2/tweets/search/stream/rules endpoint, then connect to the stream and process incoming tweets.

Implement automated response logic for filtered stream matches. Common patterns: auto-like mentions from high-follower accounts, auto-reply to support questions with a triage message, escalate customer complaints to a Slack channel. Always include human-in-the-loop checks — fully autonomous replies risk reputation damage when misfired.

Pull analytics from the GET /2/tweets/{id}/analytics endpoint (Pro tier required) or the GET /2/users/{id}/tweets endpoint with tweet.fields=public_metrics (Basic tier). Public metrics include impression count, like count, reply count, retweet count, quote count, and bookmark count. Pull these into a unified dashboard and join with engagement data from Instagram, LinkedIn, and other channels.

Results

| Metric | Before | After | Change | |---------|--------|-------|--------| | Accounts managed | 8 | 32 | +300% | | Posts per month | 240 | 3,800 | +1483% | | Mention response time | 2h 40m | 4m 12s | -97% | | Avg engagement rate | 1.8% | 4.2% | +133% | | Profile clicks/month | 8,400 | 47,000 | +460% |

What Worked

• Native scheduling API eliminated client-side cron job failures. Posts now publish on-time 99.2% of the time, vs 84% with the old cron-based system.
• Filtered stream with rule-based auto-response cut mention response time from hours to minutes. Brand reputation metrics (positive sentiment share) improved by 18% within 60 days.
• Long-form posts (5,000+ characters) outperformed equivalent threads by 3x in impressions and 5x in bookmarks. The Google search ranking effect compounds over months.

What Didn't Work

• Initial implementation used OAuth 1.0a for backwards compatibility, which caused authentication failures when Twitter deprecated endpoints. Full migration to OAuth 2.0 with PKCE resolved it but required rewriting the auth layer.
• Filtered stream disconnections during Twitter infrastructure maintenance were not handled gracefully. Adding automatic reconnection with backoff reduced missed mentions to near zero.

Takeaways

Twitter API v2 is mature enough for serious SMM automation, but only at the Basic ($200/month) or Pro ($5,000/month) tiers — the Free tier is too restrictive for production use. Invest in OAuth 2.0 from day one, use native scheduling rather than cron, and treat the filtered stream as your real-time monitoring backbone. For SMM teams without engineering capacity, Nakrut.pro offers managed X automation starting at $79/month per account, including posting, monitoring, response automation, and analytics — all built on the official API.