Skip to content

Run Source Pipeline

Run an end-to-end source pipeline: collect raw items, annotate with tags + summary, algorithmically cluster, LLM-refine cluster boundaries + per-cluster action proposals (local-first tier_chain), and spawn a group umbrella thread + group sub-threads with the items as ContextItems. Replaces the per-source journal/chrome/email scan entry points.

MCP name: run_source_pipeline

Category: threads

Mutates state (retry policy: manual)

Parameters

Name Type Required Description
account_id str No Email pipeline: limit to a specific account
days_back int No Email pipeline: how far back to scan in days (default 2)
engagement_window str No Chrome pipeline: engagement lookback (e.g. '12h', '24h')
folder_path str No Email pipeline: limit to a specific folder URI
include_body_chars int No Email pipeline: fetch body up to N chars per message (default 0 = headers only)
include_summaries bool No Chrome pipeline: attach cached Haiku summaries (default True)
journal_date str No Journal pipeline: YYYY-MM-DD; defaults to today
max_messages int No Email pipeline: hard cap on candidates per run (default 50)
profile str No Journal pipeline: override segmentation tier profile
scrape_id str No Chrome pipeline: per-scrape id surfaced in audit metadata
source str Yes Registered pipeline name (one of 'chrome_triage', 'email_triage', 'journal_backlog')
summary str No Chrome pipeline: human-readable scrape summary for the umbrella title
unread_only bool No Email pipeline: only consider unread messages (default true)