Nebula fits the stack
you already run on.
Capture from any conversation, then execute across your CRM, docs, tickets, and comms — with least-privilege, per-tool write access you approve.
No integration required.
Nebula captures at the operating-system level, so it works with any meeting platform — and with conversations that aren't in a meeting platform at all.
One model, trained for this job.
Nebula runs on our own model, trained on how conversations turn into work — not a general-purpose chatbot pointed at a transcript. The same engine drives every stage, so what it hears, what it plans and what it executes stay consistent.
Hears the room
Transcription and speaker separation tuned for real meetings — crosstalk, accents, product names and the half-finished sentences people actually speak.
Plans the work
Trained specifically on the step everything else skips: turning what was said into concrete, executable actions with the right owner, target and detail.
Checked before it runs
The model proposes; a symbolic layer decides. Every action is tested for grounding, completeness and authorization before anything reaches your tools.
One engine end to end
Capture, recaps, planning, verification and execution all run on the same model. No hand-offs between vendors, no behaviour that changes because a provider shipped an update.
Improves on our data, not yours
The engine is trained on data we own and license. Your conversations are never training material — on any plan, at any time.
Run it where you need it
Pin the engine to a region, or deploy it inside your own network so no conversation leaves your infrastructure. Available on Enterprise for teams with strict data rules.
Where the work actually lands.
Write access is scoped per tool and per action. Items marked beta ship with the execution agent as design partners come online.
CRM & sales
Tickets & projects
Docs & knowledge
Comms & calendar
Storage & automation
Build on top of every conversation.
A REST API and webhooks give you programmatic access to transcripts, summaries, actions and analytics — so Nebula becomes a data source for your own tools. Rolling out with the execution agent.
- REST API with scoped access tokens
- Real-time webhooks on every event
- SDKs for JavaScript & Python
Docs aren't public yet. Ask for early API access
// React the moment an action is executed const nebula = require("@nebula/sdk"); nebula.on("action.executed", async (evt) => { const { type, target, source } = evt.data; await audit.record({ action: type, utterance: source.utterance, at: source.timestamp, }); });
Before you connect anything.
Its own. The Nebula engine is trained specifically on how conversations turn into work, rather than adapted from a general-purpose chatbot, and the same model drives transcription, planning, verification and execution. It is included in every plan — no separate AI subscription, no per-token bill. Enterprise can pin it to a region or run it self-hosted.
As little as possible. Write access is requested per tool and per action rather than granted wholesale — permission to move a deal stage, not permission to your whole CRM. You approve each scope, and external or irreversible actions require approval at execution time by default.
Yes — and it doesn't need a per-platform bot to do it. Nebula captures at the operating-system level on Windows and macOS, so it records any meeting platform, any browser tab, a phone call on speaker, or an in-person conversation, including apps with no API at all.
Don't see your tool?
Tell us what you need.
Design partners get their three most-used tools wired up first.