Watch keywords
Starts a run from collected posts matching terms, cashtags or phrases, with literal include/exclude filters. Missing collector coverage is unavailable, not zero activity.
Node ID twitter-watch-keywordsMarkdown
How it works
Starts runs from newly collected posts matching any configured term, cashtag or phrase. Matching is literal and case-insensitive, not executable search syntax.
Include/exclude terms further narrow the sampled events. Use an enricher and decider when you need token evidence rather than immediately treating a post as a reason to trade.
Settings
Initial values for a new node, not your saved graph. Existing graphs can have different settings.
| Setting | Initial value | Details |
|---|---|---|
Terms / cashtags / phrasesterms | None | Comma-separated literal terms searched within collected public posts, not unrestricted X coverage. |
Include termsinclude | None | Optional additional literal include terms. |
Exclude termsexclude | None | Matching any excluded term rejects the post. |
Maximum post agemaximumAgeSeconds | 900 | Editor range: 15 to 3600. |
Connections
These lists show compatible node types, not approval of every graph. Branch, position context, subscription, no-cycle and complete-exit rules also apply. A node cannot connect to itself.
Before you run
- Requires Tier 2. Running graphs request bounded literal-phrase searches. Terms may use letters, numbers, spaces, $, @, _, dots and hyphens; other punctuation is local-only and cannot widen the upstream query. Include/exclude apply locally. Quotas, collection delay and rotating capacity apply.
- Up to 24 subscriptions per graph, 48 distinct per account and 96 shared requested feeds. The collector keeps its maximum 20 account and eight keyword feeds including baseline sources; excess feeds rotate every 15 minutes. Pausing/deleting releases subscriptions, and stale leases expire in 90 seconds. No historical backfill or fabricated token association.