Where you plug in
Connect a source
Each Lookup source is a fetcher agent over a real backend — Postgres, InfluxDB, Redis, Snowflake, or a local file. The registry holds only the schema; the agent pulls a filtered slice on demand. Live status below; upload a CSV to add your own data.
Upload
Upload CSV
liveAny CSV becomes a queryable source — headers become columns, types are inferred automatically. Max 50 MB / 50 000 rows.
drop a CSV here · or click to browse
Live sources · fetcher agents
plant_erp
ERP — suppliers, materials, receipts
3 object(s) reachable
plant_mes
MES / SFC — machines, work orders, lots
4 object(s) reachable
plant_lims
LIMS / QMS — inspections
1 object(s) reachable
plant_cmms
CMMS — maintenance events
1 object(s) reachable
plant_historian
Historian — machine sensor time-series
1 object(s) reachable
adventureworks
AdventureWorks — Production · Purchasing · Sales
16 object(s) reachable
fab_mes
fab_mes
3 object(s) reachable
fab_eqp
fab_eqp
1 object(s) reachable
fab_fdc
fab_fdc
1 object(s) reachable
fab_metro
fab_metro
2 object(s) reachable
fab_etest
fab_etest
1 object(s) reachable
fab_sort
fab_sort
1 object(s) reachable
snowflake
Snowflake — analytics warehouse (daily/weekly/quarterly roll-ups)
ENOENT: no such file or directory, scandir '/app/data/lookup/sources/snowflake'
csv_uploads
CSV Uploads — ad-hoc files promoted to queryable sources
no uploads yet — upload a CSV on /connect
mssql
mssql
ENOENT: no such file or directory, scandir '/app/data/lookup/sources/mssql'
rest
rest
ENOENT: no such file or directory, scandir '/app/data/lookup/sources/rest'
s3
s3
ENOENT: no such file or directory, scandir '/app/data/lookup/sources/s3'
Connector catalog
SECOM
livefile · semiconductor
1,567 wafer lots × 590 process sensors. The messy hero — drives the bad-data refusal.
AI4I 2020
livefile · machine telemetry
10,000 machines, quality L/M/H, 5 failure modes. The clean contrast.
NASA C-MAPSS
livefile · turbofan RUL
Run-to-failure trajectories, 21 sensors. Powers asset reliability.
InfluxDB
liverest · time-series
InfluxDB v2 historian via Flux API. Set INFLUXDB_URL/TOKEN/ORG/BUCKET to activate; falls back to Redis → file.
CSV Upload
livefile · ad-hoc
Upload any CSV; headers become columns, types are inferred. Immediately queryable via the stitch engine.
Snowflake
livesql · warehouse
Daily/weekly/quarterly roll-ups from Snowflake. supplier_id cross-joins back to ERP. Set SNOWFLAKE_ACCOUNT/USER/PASSWORD/WAREHOUSE/DATABASE.
MES / SFC
plannedstream
Manufacturing execution / shop-floor control events — work orders, routings, scrap.
OPC-UA native
plannedstream
Process historian via a native OPC-UA gateway — tag time-series at line rate.
MQTT (HiveMQ/EMQX)
plannedstream
MQTT broker integration for real-time sensor events — HiveMQ cloud or self-hosted EMQX.
S3 / Azure Blob
plannedobject store
Parquet/CSV files in cloud object storage — same CSV adapter, async ingestion path.
LIMS / QMS
plannedapi
Lab & quality management — test results, CoAs, dispositions.
ERP (live)
plannedapi
Orders, BOMs, inventory, supplier master — the sourcing-policy backbone.
CMMS (live)
plannedapi
Maintenance management — work orders, asset health, downtime.