Runs on your servers · on-premise · air-gapped if you need it
NimbleMeet is a self-hosted video conferencing platform built on modern media protocols. It writes your meeting minutes automatically, keeps server load to a minimum, and fits into the infrastructure you already run — including your existing fleet of SIP room systems. Deploys in a single business day.
SFU efficiency for the web and MCU compatibility for your hardware fleet — in one system. Migrate without replacing equipment and without retraining users.
SFU for the web, MCU for hardware room systems. Both modes run in the same conference at the same time — no compromises, no either-or.
When a conference has only browser participants, NimbleMeet runs as a pure SFU: one stream in from each participant, targeted forwarding to everyone else, no intermediate transcoding. AV1 and VP9 run at full strength while the server stays nearly idle.
The moment a hardware endpoint or an external conferencing server joins, the platform automatically renders a server-side composite for it — a ready-made layout in the codec that device expects. Web participants keep flowing through the SFU, so load does not grow proportionally.
Legacy solutions either save resources (SFU with no hardware support) or work with any device (MCU with full transcoding). NimbleMeet does both within a single conference — the mode is chosen automatically per participant type.
The traditional approach: the server decodes every stream and encodes a shared composite. Reliable and familiar, but resource-hungry — server load grows sharply with meeting size.
The server receives a single stream from each participant and routes it to the others at the right quality. Transcoding kicks in only where it is genuinely needed — for example, when hardware SIP endpoints join.
After every meeting, NimbleMeet automatically produces structured minutes: a short summary, topics with timestamps, decisions made, action items with owners and due dates, and a full transcript — in Markdown, PDF, or plain text.
Speech recognition is powered by NimbleGPT, our proprietary multilingual model trained for real meeting acoustics: speakerphones, SIP telephony, and mobile connections. It runs strictly on-premise — not a single audio frame or line of transcript ever leaves your perimeter.
NimbleGPT — a product in its own right Capabilities, deployment options, and technical details on a dedicated pageHardware room systems, external conferencing servers, corporate directories, and standard Linux — everything connects without replacing your infrastructure.
Poly (G7500, X30/X50, VVX), Yealink, Cisco, Grandstream, and other SIP endpoints.
A partner conferencing server joins your layout as a single tile. Dial out to an external server or endpoint by URI or address.
Outbound calls route by numeric prefix to the right gateway or trunk, with a route preview right in the interface.
Each device type gets its own codec and media profile. Hardware endpoints join cleanly without per-call manual setup.
LDAP / Active Directory. SAML/OIDC on the roadmap.
iCal invitations (.ics). Exchange EWS and CalDAV are planned.
Recent Chrome, Edge, Firefox, and Safari. No plugins.
Ubuntu 22.04 / 24.04 LTS. Other Linux distributions on request.
x86_64 (primary), ARM64 (available).
RTSP from IP cameras, RTMP/WHIP ingest for external studio feeds.
MP4 recording to internal storage, outbound RTMP streaming, HLS playback.
Upload a video to the internal library and play it into a room as a participant — for venue checks or presenting material.
An isolated perimeter, transparent auditing, role-based access, and modern cryptography — built into the platform, not sold as an add-on module.
The roadmap is open. Priorities are shaped by requests from enterprise customers.
Extended role model and delegation, SAML/OIDC single sign-on, SCIM user provisioning.
Breakout rooms, webinar mode for large broadcasts, PSTN dial-in to conferences.
Deeper integrations with corporate systems (Exchange, Google Workspace, Slack), expanded NimbleGPT language coverage.
Server-side ML noise suppression, live speech translation, ISO 27001 / SOC 2 compliance program.
We'll prepare a demo tailored to your environment, walk through integration with your existing infrastructure, and propose a licensing model.
Real people, no automated drip campaigns. We usually reply within one business day.