Navaday Partner Pipeline

Updated as of 10/13/2026

Engineering, CAD and manufacturing

  1. Large-scale CAD archive: Approximately 2.2 million 2D and 3D CAD files spanning mechanical engineering, aerospace and architecture. Sourced through a data aggregator; sublicensing rights and provenance require verification.
  2. Architecture project histories: Approximately 1,650 projects spanning 20 years, from initial sketches to completed buildings, with milestones and costs.
  3. Precision manufacturing CAD and production trace data: Approximately 10 years of operations producing automotive, defense and aerospace parts and dies. Includes CAD models, CAM machining files, 2D drawings for every manufactured part, and G-code/NC-code containing cutting coordinates. Per-part metadata includes the machine used, material, location and part cost. File counts have not been specified.
  4. Fabrication and dimensional verification data: LiDAR point clouds, as-built models, dimensional verification data, fabrication and shop drawings, CAD models and weld documentation. Includes work instructions with revision histories, inspection records and production records.
  5. Medical-device engineering and manufacturing data: More than 15 years of CAD drawings and engineering models, assembly, inspection and repair videos, SOPs and engineering/manufacturing process documentation. Approximately 95% English. The same archive includes 40,000 spine X-rays and approximately 30,000 annotated motion X-ray videos.
  6. Facility engineering and technical documentation: Decades of documentation for airports, stadiums and other complex facilities, including technical and network architecture, systems diagrams, plans, renderings and workflows.
  7. Commercial drone and geospatial data: Approximately 5 TB from nearly a decade of client work. Includes construction-progress imagery (307 GB), thermal/infrared imagery (141 GB), LiDAR/survey data (167 GB), offshore-energy data (79 GB) and oil-and-gas data (52 GB), plus point clouds, 3D meshes and CAD files in DWG, DXF, RVT and SKP formats.
  8. Construction project records: More than 15 years and 250 projects, including residential developments exceeding $100 million. Approximately 1 TB of daily logs, RFIs, change orders, budgets, tenders, schedules and takeoffs.
  9. Aviation and aerial firefighting operations: Approximately 18 years of operational data across multiple countries and agencies, containing 1.5 billion historical records and growing by more than 1 million records daily.
  10. Commerce, manufacturing and finance workflows: Operating data from a direct-to-consumer and SaaS business with factories in China and Mexico and approximately 600–700 employees. Includes workflow, manufacturing and finance systems.

Agents and safety

  1. Workforce activity data: Keyboard, mouse and application activity from approximately 250,000 workers across 10,000 companies and 50 job types, with up to five years of history.
  2. Social engineering and attack data: 110,000 recorded human phishing calls over 12 years, approximately 34,000 usable, labeled by pretext, outcome and influence tactics. Also hundreds of thousands of phishing and SMS attack records.

Medical

  1. Clinical conversations and patient records: 50,000–60,000 hours of doctor-patient audio, 2.2 million transcripts and 65 million patient records with clinician notes.
  2. Consultation audio with clinical documentation: Approximately 150,000 hours annually of real consultation audio paired with verbatim transcripts and signed doctor notes. Approximately 90% Swedish.
  3. Long-horizon clinical task data: 50 million clinician-verified data points.
  4. Radiology archive: Approximately 150 million DICOM images totaling 80–90 TB, including 10 years of mammography paired with reports.
  5. Virtual inpatient consultation records: Clinician documentation from approximately 25,000 consultations monthly and more than 220,000 annually, across over 400 programs in all 50 states, tied to health-system EHR workflows.
  6. Ultrasound and longitudinal patient records: Approximately 6,000 thyroid and abdominal ultrasound studies containing 20,000 images and totaling 400 GB. Studies from roughly the past year include interpretation reports. Also approximately 18,000 active patients, 24,000 encounters since December, 10 years of patient records, insurance claims, glucose-monitor reports, and patient calls and texts.

Video, visual and creative production

  1. Commercial video and production trace archive: Approximately 2.07 PB containing 2.53 million video files, including 931 TB of camera-original footage. The same productions are preserved across raw footage, graded masters and final cuts, alongside 1.47 million project files containing briefs, scripts, storyboards and timelines. Also includes 1.19 million RAW stills and 448,000 uncompressed audio files, with voiceover, location sound and music separated. Approximately 3% of the archive was previously licensed to an AI lab.
  2. Commercial production archive: More than 500 TB from thousands of productions spanning 2020–2026. Approximately 55–65% video, 25–30% RAW photography and product/lifestyle imagery, and 8–12% SOPs, creative briefs and shot lists. Rights to license the full corpus are confirmed.
  3. Live-action, animation and production assets: Approximately 100 TB of raw and edited live-action footage, plus animation, 3D models, production-ready rigs, motion capture, concept art and playblasts with review feedback and corrections. Includes approximately 500 GB of code version history, 3.9 TB of production version-control data and original music.
  4. Music and stems: 6,000 rights-owned music tracks and 45,000 stems, manually tagged by genre, mood, BPM and key.

Driving and physical operations

  1. Driving video and fleet telemetry: Vehicle-mounted road video covering approximately 2 million miles monthly, paired with speed, braking and G-force telemetry. An additional fleet dataset contains approximately 0.5 PB of telemetry spanning 11 years.
  2. Recycling pickup imagery: Tens of thousands of photos monthly from recycling pickups across the US, supported by 13 years of operational history.
  3. Agronomy and agricultural commerce data: More than 100,000 crop, leaf and plant-disease images; 1.16 million product-use records linked to farm coordinates; and approximately 2,300 hours of Hindi support calls from a recent 30-day period. Also approximately 190,000 orders, 96,000 credit transactions and 530,000 inventory and pricing records.

Conversations, calls and audio

  1. Creator and customer conversations: 321 million text conversations spanning more than 10 years, including voice notes and photos.
  2. Technical interviews: Over 500,000 hours of recorded interviews, paired with interviewer notes.
  3. Healthcare front-desk conversations: Approximately 100 million de-identified text and voice conversations from small and midsize US healthcare practices, growing by approximately 350,000 monthly. Covers scheduling, payments, FAQs, waitlists and reminders; approximately 25% are complex multi-turn cases.
  4. Brazilian sales and support interactions: Approximately 1 million customer interactions monthly, 70% written and 30% voice. All calls recorded with speech analytics and quality assessments. Approximately 60% sales and 40% support, in Brazilian Portuguese.
  5. Multilingual call-center audio: Approximately 500 agents handling more than 100,000 calls daily in English, Hindi and regional languages, with approximately 5 TB of audio.
  6. Annotated coaching conversations and business workflows: 59,000 hours of coaching conversations spanning approximately 15 years, annotated with speaker labels, call-flow maps, emotion labels, coaching-move labels and voice measures. Also approximately 1,000 hours of internal team calls, 1.79 TB of cloud-drive data with edit history back to 2012, and 54 GB of company email. Twenty sample call recordings have already been shared.
  7. Chef marketplace interactions and imagery: Customer, chef and support call recordings and chats from a marketplace with approximately 1.8 million registered users. Includes approximately 100 GB of chef check-in and check-out photos showing kitchens and prepared food, plus approximately 10 code repositories.
  8. Messaging, sales calls and CRM records: Approximately 2–2.5 TB of conversational data, primarily from approximately 5,000 messaging groups. Includes recorded sales and customer-success calls from the past 1.5 years and CRM data from 2022–2026.
  9. Sales conversations, design assets and internal workflows: Sales-call audio and transcripts, eight years of human-created graphics, four to five years of internal chat history and a 1.3 TB cloud drive. Entirely English.

Code and product

  1. Graded coding and interview data: 67,255 graded code attempts containing 2.5 million lines of user-generated code; 42,162 project submissions across 199 assignments; 16,925 AI interviews from 2024, including 10,208 with grader output; and 37,970 data-structures-and-algorithms attempts. Covers 101,037 unique people and includes 466,325 community-chat messages.

Finance and deals

  1. M&A transaction records: 50,000 completed deals with full buyer-seller messaging and due diligence.
  2. SPV and fund administration archive: Records from 2015–2022 covering subscriptions, KYC/AML, closings, capital calls, wires, cap tables and investor communications. Includes more than 1 million document-storage files, over 770 GB of cloud-drive data and over 200 GB of email.
  3. Institutional real estate transaction data: More than 6,000 acquisition interactions spanning approximately three years, including submissions, acceptances, rejections, counteroffers, negotiations and underwriting questions. Supporting documents include purchase contracts, addenda, settlement statements, inspection reports and property-condition reports.

Retail and research

  1. Specialty retail operating records: Full operating history of a US healthcare and work-apparel retailer with 35 stores across 11 states, acquired out of bankruptcy.
  2. Market research and forecasts: Market sizing and forecasts covering thousands of markets globally, produced by approximately 250 research staff.

Code

  1. Private code corpus and software engineering histories: More than 22,000 repositories spanning backend, frontend, data engineering, infrastructure, AI/ML and other software domains. Includes source code, Git history, commits and contributor records, with pull requests, issues, review discussions, tests, CI configurations and technical documentation where available. Covers multiple languages, frameworks and industries.

Enterprise Workflows

  1. Enterprise workflows and business systems: Data from more than 50 organizations across logistics, fintech, manufacturing and health technology, with source companies ranging from approximately 200 to 800+ employees. Includes internal messaging, business-system records and associated codebases. Sourced through a data aggregator; some datasets have already been licensed exclusively, so availability and sublicensing rights require confirmation.