Wenti Labs
Wenti Labs builds AI agents that turn unstructured site activity into structured construction documentation. The agents pull from project photos, emails, and chat messages across WhatsApp and Telegram, then generate progress reports, safety records, defect logs, and customer quotations. Wenti Labs says its system uses retrieval-augmented generation anchored to a defined set of company data sources to reduce hallucinations and keep outputs grounded in project records.
Best for: Contractors, developers, and surveyors who want to automate site reporting, safety docs, and QA/QC from photos and messages.
Last reviewed 6/2/26
- Pricing
- Free plan, Pro plan (unlimited use cases, file uploads, multiple users), and Enterprise plan (on-premise LLM deployment, SSO). Per-tier dollar pricing not published on their site.
- Integrations
- WhatsApp, Telegram, Google Drive, Microsoft Word, Microsoft Excel, PDF uploads
- Property types
- Commercial construction, Infrastructure, Built environment
- Geography served
- Singapore, with stated expansion plans into Japan and Australia
How Wenti Labs uses AI[1]
Wenti Labs builds AI agents that turn unstructured site activity into structured construction documentation. The agents pull from project photos, emails, and chat messages across WhatsApp and Telegram, then generate progress reports, safety records, defect logs, and customer quotations. Wenti Labs says its system uses retrieval-augmented generation anchored to a defined set of company data sources to reduce hallucinations and keep outputs grounded in project records.
- • AI agents convert site photos, emails, and chat threads into structured progress, safety, and QA/QC reports
- • Site Capture documents defects from photos and compiles them into reports
- • Retrieval-augmented generation grounds outputs in a defined set of company data sources to limit hallucinations, per Wenti Labs
- • Generates customer quotations from document analysis and pushes updates back into WhatsApp and Telegram
- • Enterprise plan offers on-premise LLM deployment and SSO
AI type: RAG-based AI agents for construction documentation
API: unknown MCP: unknown
Recognition[2]
Key numbers[1][2][3]
- • Over 20 paying enterprise clients
- • approaching US$1M in annual recurring revenue and reported profitable with roughly 40% month-over-month growth
- • generating recurring revenue from enterprise customers before raising external capital
- • deployed with Tier-1 Singapore contractors including Boustead Projects E&C, Penta-Ocean Construction, and Woh Hup
- • partner of NVIDIA, AWS, and Microsoft for Startups programs
Credibility[1][2][3]
- Founders
- Founded in 2023 by co-founders Ethan Ow (CEO), Nguyen Tu, Gong Suk Yee, and Kwok Yangbin. Ethan Ow previously worked as a construction project manager at CapitaLand and held operations roles at Uber and AWS; the founding team is described as serial entrepreneurs with experience across construction site operations, enterprise software, and applied AI.
- Customers
- Deployed on active construction projects with Boustead Projects E&C, Penta-Ocean Construction, Woh Hup, and multiple Tier-1 general contractors in Singapore. Reported over 20 paying enterprise clients.
- Investors
- Pre-seed round led by Zacua Ventures, with participation from Aurum Investments and Feedback Ventures.
Founded
2023
Headquarters
Singapore
Stage
Pre-seed
Related companies
Civils.ai
AI takeoff and document-review platform for construction contractors and consultants that pulls quantities and geotechnical data from PDF drawings and checks them against specs and codes.
AI agents read PDF drawings and reports, extract takeoff quantities (areas, lengths, volumes, counts) and subsurface geotechnical data, and run compliance checks against specs and codes with cited sources, all under human QA review. The vendor claims 90% less manual effort and 97% accuracy on modern PDFs.
Doxel
Doxel is an AI construction progress-tracking platform that turns site scans into objective work-in-place data, helping owners and general contractors keep projects on schedule and on budget.
Doxel uses computer vision and 3D semantic deep learning to compare reality-capture scans of a job site against the BIM model and schedule. It measures completed work by component and trade, then flags schedule and budget risks before they grow.
ClearEdge3D
Converts 3D point cloud scans into BIM-ready models using computer vision, automating as-built modeling and construction QA/QC with claimed 73% time savings.
Computer vision and pattern-matching algorithms extract building components from 3D laser-scan point clouds and fit spec-based geometry to auto-generate BIM (Building Information Modeling) models, then compare as-built scans against design models to flag deviations.