Data platform · AI-native · Public cloud · Private cloud · On-premises
Konobase is a data and AI platform for researchers, teams, and businesses that demand full ownership of their data. Store, access, and analyze on our cloud, your cloud, or your own hardware — with AI that runs where your data lives, never somewhere else.
Deploy your way
Fully managed by Konobase. Zero infrastructure to run. SLA-backed uptime, automatic updates, and instant provisioning. Start in seconds.
Konobase deployed into your own AWS, Azure, or GCP account. You own the infrastructure and the data. We manage the software.
A single binary on your own hardware. Air-gapped networks, classified environments, or data-residency mandates — if you have a server, Konobase runs on it.
What's inside
A full-featured data platform that works with the tools you already use.
Your shared file system — for a lab, a team, or just yourself.
Upload, organise, move, rename, clone, trash, and restore files and folders. Nested hierarchy with per-item permission control.
Publish datasets, reports, or portfolios with a direct link.
Make any folder public with a branded URL at /u/username/public/. Share individual files with direct links. Inline preview for images, video, audio, and PDFs.
Keep lab notes, documentation, or meeting notes next to your data.
Create, edit, and store notes directly in your file hierarchy. Rich markdown editing with live persistence — no separate app required.
Know exactly what's stored and who's using what.
Real-time storage usage shown in the sidebar. Per-user quotas, file count, and a clean storage dashboard in settings. Zero operational overhead.
Find any image by what it looks like, not what it's named.
Automatic thumbnails for all images. Perceptual fingerprinting lets you find visually similar photos across your entire library in one click.
Record field audio, lectures, or meetings — stored right alongside your files.
Built-in music player with queue and shuffle. Record voice memos or video directly from the browser — no plugins. Playback with waveform preview in the details panel.
Version-control code and data together, in one authenticated place.
Create bare Git repositories inside your file space. Push code alongside documents and data — everything in one authenticated store.
Access your files from the shell without opening a new window.
Full browser-based terminal available from within the file browser. Run scripts, manage files from the shell, or SSH to another host — all in the same tab.
Access your data any way you want
Mount as a drive, use the AWS CLI, or SSH in. Your data works with any tool that speaks standard protocols.
Drop-in replacement for Amazon S3. Use the AWS CLI, boto3, or any S3 SDK with your existing code — just point the endpoint at Konobase.
Mount your Konobase as a network drive on macOS, Windows, or Linux. Drag and drop files in Finder or Explorer — no client software needed.
Secure FTP on port 2222 using your WebDAV credentials. Works with FileZilla, Cyberduck, rsync, and any other SFTP client or automation script.
Zero-trust access model
Every authentication method, every access, every change — logged and queryable.
WebAuthn passkeys are the default. No passwords to steal, no 2FA codes to intercept. Biometric login from any device that supports FIDO2.
Every file operation — upload, download, delete, rename, share — is timestamped and attributed to a user. Queryable by admins. Required for HIPAA, FERPA, and SOC 2.
Set read/write permissions per file or folder. Public folders with direct links, private folders locked to the owner, or shared via authenticated WebDAV credentials.
On private cloud and on-prem deployments, your data never touches our systems. Run on bare metal, Kubernetes, or a single VM. Embedded storage keeps ops simple; continuous replication keeps it safe.
Built AI-first — not AI-appended.
Most platforms bolt a chat interface onto existing file storage. Konobase is different: the storage model, permission system, and access layer were designed from day one knowing AI would read, classify, and reason over your data. The capabilities below are built and feature-flagged — they go live in days.
Ask questions in plain English. Find documents by meaning, not just filename. "Show me contracts from last quarter mentioning the NDA clause."
Ask questions about the contents of any file — PDFs, Word docs, spreadsheets, code. Get answers with source citations, not hallucinations.
Upload a batch of files and let the AI suggest folder structures, tags, and descriptive names based on content — not just file extension.
Voice memos and video recordings are automatically transcribed, searchable, and summarised. Every spoken word indexed alongside your documents.
Beyond perceptual hashing — deep embedding models that find visually similar images even with different cropping, colour grading, or resolution.
AI runs on your infrastructure. Self-hosted models (Ollama, vLLM) or your own API keys. Your data never leaves your environment to train any third-party model.