Building initial product · pilots opening
Kilnvera reads your kiln’s firing logs and inspection counts, scores every batch for quality risk, and shows production managers which part of the curve went wrong: too fast through quartz inversion, a short soak, an overshoot. Then it says what to change.
For brick, tile, sanitaryware, tableware and refractory plants.
Batch B-2411 · Kiln K1
Floor tile 600x600
Fast cooling through quartz inversion
Cooled at up to 197 °C/h between 650 and 500 °C (limit 100 °C/h).
Inspected defect rate for this batch: 13.5%
Output of the Kilnvera engine on our fictional sample firing log.
0–100 risk score
Every batch in the log gets an interpretable quality-risk score with the evidence behind it.
Seven firing-curve checks
Ramp rates in critical zones, soak time, peak temperature and setpoint tracking.
Defect correlation
Links each check to your inspection counts to show which faults cost you the most ware.
Fuel per kg screening
Flags batches that burned far more fuel per kilogram than the rest of the dataset.
The problem
In a brick or tile plant, a rejected batch has usually already been paid for twice: in raw material and in fuel. The evidence of what went wrong was in the firing log all along.
How Kilnvera fixes it
Upload the firing-log export you already have. Kilnvera measures every segment of every batch against the limits you set, scores the risk, compares the result with your inspection counts and writes a specific corrective action. Matching dozens of batches against their curves, which rarely happens by hand, runs in seconds.
Controllers, chart recorders and data loggers record every reading, but the exports are rarely opened unless something has already gone badly wrong.
Cracks, warping and colour variation show up at sorting, 12 to 48 hours after the firing that caused them. By then the same profile has run again.
Inspection counts live in a quality sheet, firing logs on a controller. Matching them by hand across dozens of batches does not happen on a busy shift.
A cooling fan running too hard or a soak cut short can repeat for weeks, costing fired ware, fuel and kiln time on every cycle.
Core features
Every feature below works in the current workspace. Each check measures something a kiln engineer would recognise, with limits you can tune to your body, ware thickness and kiln.
Each batch gets a 0–100 score built from the checks it breached and by how much. High-risk batches rise to the top of the queue.
Decide which batches to pull for extra inspection before they ship.
Batch queue
Sample Data| B-2411 | Fast cooling through quartz inversion | 69high |
| B-2405 | Fast cooling through quartz inversion | 68high |
| B-2419 | Fast cooling through quartz inversion | 68high |
| B-2413 | Soak at top temperature cut short | 64high |
Kilnvera measures ramp rates over 15-minute windows inside the critical zones, the real hold at top temperature, peak versus setpoint and time spent off profile.
Catch the segment that causes dunting, blow-outs or under-firing.
When inspection counts are included, Kilnvera compares defect rates in batches that breached each check against those that did not, and reports correlations.
Know which fault is costing the most ware. Associations are labelled as such, not as proof.
Fast cooling through quartz inversion
Sample DataEvery finding comes with the measured value, the limit it broke and a concrete adjustment for the kiln crew.
Turn a chart into an action the shift can take.
Recommended action
Slow the cooling segment from 650 °C to 500 °C to 100 °C/h or less: reduce cooling-air fan speed or damper opening in that zone and check for doors or ports opened early.
Sample DataUsing fuel and load columns, Kilnvera computes fuel per kilogram for each batch and flags statistical outliers against the rest of the upload.
Spot burner, seal or loading problems that waste gas.
m³ per kg
Sample DataB-2417: 0.161 vs median 0.111
Open any batch to see its actual temperature against setpoint with the critical zones shaded, and hover to read exact values.
Show the crew exactly where the curve left the profile.
Every analysis is stored with its original file. Reopen it later, download the source, export per-batch results to CSV, or re-run with thresholds tuned to your plant.
Build a record of firings you can discuss with your process team.
Saved analyses
Reopen any past upload
Your thresholds
Set limits per analysis
CSV export
Per-batch metrics and risk
How it works
Step 1
Export readings from your kiln controller, SCADA historian or data logger as CSV: batch, timestamp, temperature and setpoint. Fuel, load and inspection counts are optional.
batch_id, timestamp, temp_c, setpoint_c
Step 2
The server checks the columns, skips unreadable rows and tells you how many, then measures each batch: ramps, soak, peak, tracking and fuel per kg.
4,713 rows → 24 batches
Step 3
Each measurement is compared with your limits. Breaches become findings with evidence, a likely defect type and a risk score. Inspection counts are correlated across batches.
QUARTZ_COOL · 197 °C/h > 100
Step 4
Review the queue, open the curve of any batch, export results and hand the recommendation to the kiln crew. The analysis stays saved for the next review.
export.csv · saved to history
The workspace
The analysis workspace runs end to end in our test environment: firing logs are uploaded, analysed on the server, saved and reopened. We are preparing hosted pilot workspaces and have no paying customers yet.
The workspace needs the Kilnvera analysis service. Where it is not connected, the page says so instead of showing made-up results.
Working today
In progress
Who it’s for
Production managers, kiln supervisors and quality leads in plants that fire clay and ceramic bodies.
Technology
Today the Kilnvera engine runs as a Node.js service with a local database in our test environment. The same code already includes an AWS Lambda handler and a DynamoDB and S3 storage adapter, and this is the architecture we are moving pilot workspaces onto. The predictive models on Amazon SageMaker AI and the briefings on Amazon Bedrock are the next layer, built on top of the deterministic checks rather than in place of them.
Data collection
Amazon API Gateway
Receives firing-log uploads from the workspace and enforces the pilot access code.
Secure storage
Amazon S3 · Amazon DynamoDB
S3 keeps original logs and full results, encrypted. DynamoDB keeps analysis records for fast listing.
Processing
AWS Lambda
Runs the Kilnvera engine: parsing, validation, firing metrics, rule checks and risk scoring.
AI intelligence
Amazon SageMaker AI · Amazon Bedrock
Predictive defect models trained on pilot history, and plain-language briefings from computed results.
Application
Kilnvera workspace
The React workspace shows the queue, curves, findings and exports through the same API.
Status and pilots
Stage: Building initial product. We are looking for a few plants willing to share firing-log exports and inspection counts so we can tune checks against real production.
2026
Kilnvera founded by Bidemi Ade.
Nov 2026
Analysis engine and workspace working end to end in our test environment, with automated tests for upload, analysis, persistence and access control.
Next
Hosted pilot workspaces for a small group of plants, using their own firing logs.
Tell us about your kilns. We will reply by email to agree what data to share and how the pilot workspace will be set up.
Pricing
Pilot pricing in Nigerian naira. The billing period and tier limits are confirmed in your pilot agreement before anything is charged.
Starter
₦5,000
One kiln, getting started with firing-log reviews.
Growth
₦10,000
Plants running several kilns or product lines.
Enterprise
₦20,000
Multi-site groups that need onboarding support.
Team
Kilnvera is a founder-led company at the start of its journey.
Founder
Bidemi founded Kilnvera to give kiln operators and production managers a clear, evidence-backed view of what each firing did to their ware. Bidemi leads product, engineering and pilot partnerships.
ade@kilnvera.siteContact
Questions about data formats, pricing or a pilot: send a message or email us directly.