Records
17,208412
Across 10 tables
Connectors, pipeline health, access control and the settings behind the knowledgebase.
Records
17,208412
Across 10 tables
Storage Used
184 GB6%
142 GB docs · 42 GB DB
Active Users
313
8 signed in today
Ingestion Jobs Today
384
34 ok · 3 warned · 1 failed
Data Freshness
6 min12%
Median connector lag
ORES Docket System
Scraper · Playwright
All 132 active dockets and their intervenor lists.
NYS DOS Local Law Database
Scraper · HTTP
Keyword filter: solar, battery, storage, moratorium.
ProPublica Nonprofit Explorer API
REST API
Rate limited twice overnight; 14 EINs deferred.
NYS Board of Elections
Bulk CSV
Campaign finance for town board members.
Local News RSS Monitor
RSS · 84 feeds
Watertown Times, Post-Star, Finger Lakes Times +81.
Facebook Group Monitor
Headless browser
Session cookie rejected on 3 of 26 groups.
Town Minutes OCR
Tesseract + LLM cleanup
PDF minutes from 214 towns, 96% OCR confidence.
IRS 990 Bulk Download
S3 sync + XML parse
Paused until the 2024 filing season index is posted.
Municipal Hearing Calendars
iCal + scraper
Feeds the upcoming hearings board on the dashboard.
| Connector | Started | Duration | Records | Warn | Err | Status |
|---|---|---|---|---|---|---|
| Local News RSS Monitor | Today 16:42 | 38s | 214 | 0 | 0 | Success |
| ORES Docket System | Today 16:34 | 4m 12s | 91 | 2 | 0 | Warning |
| Facebook Group Monitor | Today 12:10 | 1m 06s | 0 | 0 | 3 | Failed |
| ||||||
| Municipal Hearing Calendars | Today 13:58 | 52s | 18 | 1 | 0 | Warning |
| Town Minutes OCR | Today 03:00 | 27m 41s | 63 | 4 | 0 | Warning |
| ProPublica Nonprofit Explorer API | Today 01:30 | 6m 18s | 142 | 2 | 0 | Warning |
| ORES Docket System | Today 12:34 | 3m 49s | 74 | 0 | 0 | Success |
| Local News RSS Monitor | Today 12:27 | 34s | 187 | 0 | 0 | Success |
| NYS DOS Local Law Database | Mon 02:00 | 8m 03s | 26 | 1 | 0 | Warning |
| Municipal Hearing Calendars | Yesterday 21:58 | 49s | 11 | 0 | 0 | Success |
| Facebook Group Monitor | Yesterday 18:10 | 9m 27s | 341 | 3 | 0 | Warning |
| NYS Board of Elections | May 18 04:00 | 14m 55s | 3,910 | 0 | 0 | Success |
| IRS 990 Bulk Download | May 5 06:00 | 41m 12s | 1,284 | 2 | 0 | Warning |
Duplicate entitiesHigh
Organizations sharing an EIN or a normalized name
organizations6 vs. last week
Unresolved aliasesMedium
Individual name variants awaiting a merge decision
individuals11 vs. last week
Missing county assignmentHigh
Filings with no resolvable municipality or FIPS code
filings3 vs. last week
Stale funding recordsLow
Flows whose newest supporting 990 is older than 3 years
funding_flows4 vs. last week
Low-confidence extractionsMedium
OCR or LLM extractions below the 0.80 confidence floor
filings9 vs. last week
Hearings without a project linkLow
Calendar entries the resolver could not match to a docket
hearings7 vs. last week
LLM model
Drives natural language queries and argument classification.
Embedding model
Re-indexing 6,742 documents takes about 40 minutes.
Raw document retention
How long source PDFs stay in object storage before archival.
Model-language similarity
88%
Alert when a filed local law matches model text above this score
Narrative surge
12 items
New disinformation items in a county within 24 hours
Funding flow floor
25K USD
Notify on newly traced grants at or above this amount
Hearing lead time
7 days
Days of notice before a tracked hearing triggers a briefing
Production — Web UI
read:all
created Jan 14, 2025·last used 2 min ago
LLM function-calling gateway
read:all, search:documents
created Feb 02, 2025·last used 9 min ago
Partner — CleanPath NY
read:counties, read:projects
created Mar 28, 2025·last used yesterday
Ingestion writer
write:filings, write:narratives
created Apr 11, 2025·last used 4 min ago