Export Twitter Search Results to CSV with Python

Tweet Scraper to CSV is a small Python script that runs an X query through GET /twitter and writes the public posts Desearch returned to a CSV: id, date, text, likes, reposts, replies, link and author.

The problem

A search on X is easy to read and hard to reuse. Copying posts into a sheet by hand drops dates, mixes duplicates, and keeps posts from outside the window you meant to study.

What the app does

• ORs your terms inside parentheses, and quotes any term that has a space.
• Calls GET /twitter with sort, a UTC start date, an optional end date, language and count.
• Drops duplicate ids and any post dated outside the UTC window, then sorts newest first.
• Writes posts.csv plus a snapshot of the query, how many posts were kept, and the header cost.
• By default the author column says "a user". An org account you name can be shown as itself.

How it works

1. Build the query. Terms go inside parentheses so later filters apply to the whole group.
2. Search. GET /twitter with the key in the Authorization header.
3. Clean. Duplicate ids and posts outside the date window are removed in code.
4. Write the CSV, snapshot.json, calls.jsonl and a redacted raw response.

The code

The query builder:

def build_query(terms, extra=""):
    """OR terms always go inside parentheses: ("a b" OR c) extra."""
    xs = [t if (" " not in t or t.startswith('"')) else f'"{t}"' for t in terms if t.strip()]
    if not xs:
        raise ValueError("at least one search term is required")
    q = "(" + " OR ".join(xs) + ")"
    return f"{q} {extra}".strip()

The search call:

def search(c, query, tag="search", **params):
    p = {"query": query, **{k: v for k, v in params.items() if v not in (None, "")}}
    st, body = c.get("/twitter", p, tag)
    return st, (body if st == 200 and isinstance(body, list) else [])

The folder has tweets_csv.py, core.py, offline tests and a sample run from 2026-10-15.

Endpoints

• GET /twitter

Example output

The sample run made 1 call, kept posts inside the date window, and cost $0.0075 by the X-Desearch-Cost-Usd header. date is UTC. link is https://x.com/i/status/<id>.

Run it yourself

The public examples folder linked above includes the script and a README. Install its requirements, set DESEARCH_API_KEY from console.desearch.ai, and use the run command in that README.

What it costs

The sample call returned X-Desearch-Cost-Usd: 0.0075. One call returns up to the count you set. See pricing.

Ideas to extend it

• Schedule it and append each day's CSV.
• Add a column for the query that found the post.
• Keep a second file of the ids that fell outside the window.

{
  "mcpServers": {
    "desearch": {
      "url": "https://mcp.desearch.ai/mcp",
      "headers": {
        "Authorization": "Bearer <DESEARCH_API_KEY>"
      }
    }
  }
}

Use Desearch X search to collect public posts for the terms I give you, within the date window I set. Drop duplicate ids and posts outside that window. Write a CSV with id, date, text, likes, reposts, replies, link and author. Do not add posts you did not retrieve.