How Freightglint works
Three steps from your existing TMS to a live ETA prediction your ops team can commit to. No rip-and-replace. No parallel workflow. Predictions start surfacing in under 48 hours from initial connection.
See It In ActionFrom TMS connection to live ETA prediction in under 48 hours
Connect your TMS
Freightglint connects via REST API or native adapter to McLeod, TMW, Oracle TMS, and Axele. We pull your historical BOL and delivery data — no manual export, no CSV wrangling. Initial connection typically takes less than a business day for supported TMS platforms.
We index your lane history against the broader dataset
Your BOL delivery timestamps are indexed into the lane intelligence model alongside Freightglint's broader carrier dataset. We enrich with NOAA historical weather and live ELD stream data where available. Within 48 hours your top carrier-lane pairs have scoring and confidence bands.
Predictions surface inside your existing workflow
Predictions appear in your TMS via the integration layer or in the Freightglint dashboard. For each tendered load: predicted delivery window, confidence %, and specific risk flags (weather, carrier OTP deviation, holiday proximity). Webhook callbacks push ETA change alerts when conditions shift — before WISMO calls start.
What feeds the model
Four data streams combined in real time to produce each ETA prediction.
ELD Stream
Lane History DB
NOAA Weather Feed
User TMS Sync
How we measure accuracy — and what we don't claim
Freightglint measures ETA accuracy as the percentage of loads where the actual delivery time falls within the predicted window. Our current early-customer dataset shows approximately 91% of loads arriving within a 2-hour window centered on the predicted ETA mid-point.
This figure varies by lane density — high-density lanes (where we have 500+ historical loads) perform significantly better than thin lanes. We don't claim uniform accuracy across all corridors. When we surface a prediction for a thin lane, the confidence band is wider to reflect that uncertainty.
We don't use carrier-self-reported OTP as a primary input — we treat it as a secondary signal at best, because it's systematically biased. Our primary source is actual BOL delivery timestamps from your TMS and from our broader dataset.
Request DemoSee it run on your actual freight data
30-minute demo. Bring your 3 highest-volume carrier-lane pairs and we'll show you the scoring live — confidence bands, risk flags, and all.