Twitter Advanced Search Builder

Twitter advanced search operators

Every operator the builder writes, plus the ones worth typing by hand. They go straight into the X search box, into the advanced-search URL and into the query parameter of our advanced search API unchanged. The set is cross-checked against the community-maintained igorbrigadir/twitter-advanced-search reference; X doesn't publish an official list for web search. For a deeper, code-per-operator lookup see the search operators API reference.

What we checked against live results

Operator lists online are often years old, so we ran two of the builder's presets through our own /twitter/tweet/advanced_search endpoint and read every post on the first page:

• from:NASA filter:images min_faves:1000 since:2026-10-01 until:2026-11-01 -filter:replies — 20 of 20 posts had an image, at least 1,000 likes, were not replies and were dated 1-31 October 2026 UTC (none on 1 November, confirming that until: is exclusive).
• openai filter:blue_verified lang:en -filter:replies — 20 of 20 authors had a blue check and no post was a reply.

Two operators behave differently from what older guides say: is:retweet belongs to X's paid v2 API, not to web search (use filter:nativeretweets), and filter:verified refers to legacy verification, so filter:blue_verified is the one that matches today's blue checks.

Five things that trip people up

1. The end date is exclusive. until:2026-11-01 stops at 23:59:59 UTC on 31 October. Searching one day needs two different dates. The search-by-date guide walks through time zones.
2. Retweets are hidden by default, and button retweets drop out of search after about 7-10 days.
3. Top and Latest return different sets. Top is ranked and heavily trimmed; use Latest (the builder's default, f=live in the URL) when you need everything that matches.
4. Words match more than the text. A plain keyword also matches the author's name, handle and expanded links, which is why nasa returns posts that never say "nasa".
5. Search on X needs a login, and long result lists have to be scrolled by hand. If search itself returns nothing, check the search-not-working fixes.

Run advanced search from code

The X search page is fine for a look; for monitoring, research exports or a dashboard you need the results as data. Our /twitter/tweet/advanced_search endpoint takes the same query string and returns JSON with each post's text, author, counts and media, about 20 posts per page with a cursor for the next. Based on anonymized, aggregated API usage on TwitterAPI.io, it is the most-used endpoint on the platform. Three examples (set TWITTERAPI_IO_KEY first):

1. An account's high-engagement posts on a topic over a year

import os, requests

def search(query, max_pages=10):
    tweets, params = [], {"query": query, "queryType": "Latest"}
    for _ in range(max_pages):
        r = requests.get(
            "https://api.twitterapi.io/twitter/tweet/advanced_search",
            headers={"X-API-Key": os.environ["TWITTERAPI_IO_KEY"]},
            params=params, timeout=30,
        )
        r.raise_for_status()
        page = r.json()
        tweets += page.get("tweets", [])
        if not page.get("has_next_page"):
            break
        params["cursor"] = page["next_cursor"]
    return tweets

posts = search('from:elonmusk (AI OR "machine learning") since:2025-01-01 until:2026-01-01 min_faves:10000')
for t in posts[:5]:
    print(f"{t['likeCount']:,} likes  {t['text'][:80]}")

2. Blue-check commentary on a topic, replies included

posts = search('"AI safety" filter:blue_verified filter:replies min_replies:20 lang:en', max_pages=3)
for t in posts[:10]:
    print(f"@{t['author']['userName']}: {t['text'][:100]}  ({t['replyCount']} replies)")

3. Brand mentions from the last 24 hours, for an hourly job

posts = search('"your brand" -from:yourbrand within_time:24h -filter:nativeretweets', max_pages=5)
print(len(posts), "mentions in the last 24h")

Cost for the same pull: $0.15 per 1,000 posts returned on TwitterAPI.io (pricing) versus $0.005 per post read on X's pay-per-use API (docs.x.com pricing, which notes prices may change) — $1.50 against $50 for 10,000 posts, about 33× less. The cost line under the builder's code does this sum for your own volume.

X search page, X API or TwitterAPI.io

Queries for common jobs

• Brand monitoring — "your brand" -from:yourbrand min_faves:5 -filter:nativeretweets · poll hourly; add lang:en for English only open in builder
• Competitor activity — from:competitor -filter:replies since:2026-10-01 · sort by likes to see what landed open in builder
• Sentiment research — "topic" (love OR great OR hate OR worst) min_faves:5 lang:en · feed the text to a classifier open in builder
• Archive pull — from:account since:2025-01-01 until:2026-01-01 · paginate; about $0.15 per 1,000 posts open in builder
• Expert voices — "topic" filter:blue_verified min_replies:50 lang:en · group by author open in builder
• Customer support — @yourbrand (help OR broken OR refund) -from:yourbrand · route to your support queue open in builder

Questions

Related

• Search tweets by date — dates, times and time zones step by step.
• Twitter timestamp tool — turn a date window into since_id / max_id bounds.
• Advanced search API guide — pagination, rate limits and errors.
• Twitter archive tweet finder — search an account's full history.
• Twitter (X) API overview — every endpoint and when to use it.