Research brief: a sourced crypto digest

Question

Can a scheduled workflow collect X posts and make a short, attributed crypto video without fabricating missing items?

Finding

Local runs demonstrate collection, filtering, rendering and publication. The cookie route avoids X API read charges. It depends on a working session and scraper. The output is a filtered sample, not a complete or independently verified news service.

Claim 1: collection can use a free official API

The current X pricing documentation lists $0.005 per ordinary post read. It also describes conditional credit incentives. This does not establish unlimited free collection. At 100 distinct ordinary post reads, the post-read charge would be $0.50 before other charges or credits. The older 400-post/$2 estimate describes a larger request volume, not the current default. The precise February 6 free-tier-end date has not been established from this primary page.

Workflow decision: use the configured cookie route and record the source method. Do not call the service free forever.

Claim 2: a scraper will keep working after installation

twikit issue 408 documents a March 18 transaction-code failure. Local logs also show scraper failures and later successful runs. A historical bug does not prove every current release is broken. The installed patched environment must be checked on its own merits.

Workflow decision: stop on collection failure. Preserve the error and repair the environment before retrying.

Claim 3: every covered feature is a layer 2

This claim is too broad. See crypto-layer-2-reference. Lightning channels and Ark off-chain payments differ from Monero protocol upgrades and Litecoin extension blocks. The Monero release establishes beta-test software and a scheduled October 5 stressnet fork, not mainnet activation.

Workflow decision: record each topic's mechanism and release status. Do not turn testnet posts into mainnet news.

Claim revised or dropped

Dropped: a search date filter guarantees the returned posts have the requested dates. Replacement: check each post timestamp against the saved UTC window. Revised: a quiet card proves the coin had no news. Replacement: no posts in the queried sample passed the selection rules.

Limits

Likes and followers are selection thresholds, not accuracy measurements. Keyword matches can include irrelevant content. The system does not verify every event with its original developer source. Prior first/revised-run notes are retrospective records. Preserve them without claiming they are identical to the current published edition.