Stop Manually Digitising Borehole Logs.
Upload any PDF log — any template, any contractor. Get a structured ground profile, test data, and an audit-ready annotated PDF in under three minutes.
Concerned about data security? Check out our security solution
Pages processed
2,000+
and counting — across templates and contractors
Median runtime
2m 36s
per page
Output formats
5
XLSX · AGS · CSV · JSON · annotated PDF
How it works
Three steps. One pipeline. No spreadsheet copy-paste.
Drop a PDF.
Any borehole log — scanned, born-digital, any contractor template. No pre-processing, no field mapping.
The agent reads it.
OCR detection → field-aware extraction → groundwater + tests → cross-validation. Median runtime: 2m 36s per page.
ocr · 2 pages
template · GeoConsult AU-001
ground profile · 5 layers
test data · 6 records
→ cross-validating depths…
1m 18s elapsed · 72% complete
Download what you need.
Structured data in the format your workflow expects — plus an annotated PDF for QA review.
Live demo · Visual grounding
Every number traces back to its source.
Nothing is a black box. Every extracted value is linked bi-directionally to the exact place it came from on the original PDF — so QA is a click, not a re-read.
- Bi-directional links — click table → highlight PDF, and back
- Per-cell confidence — 0–1 score on every extracted value
- Annotated PDF export, drop-ready for the client
no signup · pre-loaded log
Use cases
Where it fits in a geotech workflow.
Four shapes the same pipeline takes — from archive cleanup to live API.
“We used to budget two weeks for a junior to digitise a project’s logs. Now it’s a Friday afternoon — every extraction traces back to the source.”
Legacy archive digitisation
Decades of paper and scanned logs become searchable, structured data. Run a project's worth in a single overnight batch.
Cross-contractor log conversion
Different firm, different template — same output schema. Standardise everything before it hits your model or report.
Project data consolidation
Logs from multiple investigations consolidated into one consistent, project-level dataset. Diff between phases, compare across boreholes.
Pipeline integration via API
REST endpoints plus webhooks on job completion. Pipe extractions straight into your data lake, GIS, or modelling tool.
Security & control
We assumed the data was sensitive. We built like it.
We built BoreholeAI assuming the data is sensitive. Project files, site coordinates, client identities — none of it should be a liability.
OCR-first pipeline
Only extracted text reaches the model. The full document never leaves the OCR layer.
Auto-delete & full control
Files auto-delete after 10 days, or remove them instantly from your dashboard.
Encrypted in transit & at rest
TLS 1.3 from upload to download. AES-256 at rest. Per-tenant key isolation.
Self-hosted option (Enterprise)
Deploy in your own VPC or air-gapped network. No data ever leaves your perimeter.
SOC 2 compliant infrastructure
All storage and infrastructure providers we rely on are SOC 2 attested. GDPR-aligned. DPA available on request.
No training on your data
Your documents are never used to train models. Period. Contractually guaranteed.
Delete-on-download via SDK
EnterprisePull results through the Python SDK or REST API and the server-side file is deleted in the same call. No persistent copy after retrieval.
- Upload your PDF or image borehole log
- Our system extracts everything relevant to geotechnical engineers: ground profiles, test data, groundwater levels, and metadata
- Results are typically ready within 2–5 minutes per page, depending on complexity
- Batch processing runs multiple logs in parallel — 100 logs can complete in as little as 10 minutes
Try it with your own log.
Pre-loaded demo data. Real extraction. No credit card. 60 seconds.