Record & import
Microphone capture with chunked uploads that stay resilient on weak networks. Import existing audio and video files, with optional screen-video attachments on Web.
Self-hosted voice workspace
A voice workspace built for enterprise security.
VoxSpace records and imports audio and video, transcribes speech with AI ASR, and produces structured, traceable AI summaries. It is fully self-hosted — deployable entirely offline, connectable to your own ASR and LLM models — so voice data never has to leave your organization’s network.
No third-party cloud is required anywhere in the data path.
Vendor review — Q3 infrastructure
Synced
-
Decision: inference stays on our own cluster. No external ASR calls for this program.
-
Agreed. Rollout stays behind the VPN, and we bind phones with the QR flow only.
-
Right — I’ll draft the migration plan this week…
Live preview
AI Summary · traceable
Decisions
Inference runs on the internal cluster.09:41:07
Action items
Migration plan draft — L. Kim.09:41:26
Product
From spoken word to structured record.
Every recording becomes a session: original media, live preview, final transcript, confirmed speakers, and an AI summary that cites its sources — organized in workspaces your team controls.
Live preview, final transcript
A real-time preview while you record; an accurate post-session final transcript. Speaker separation with names you confirm, and mixed Chinese-English speech.
Structured AI summaries
Overview, decisions, action items, and key quotes — every claim traceable to transcript timestamps and citations, so nothing floats free of the record.
Workspaces & roles
Personal and shared workspaces with Owner, Admin, Editor, and Viewer roles — plus encrypted content search within a workspace.
Export anywhere
Transcripts as PDF, Markdown, TXT, SRT, or VTT. Summaries as Markdown or PDF. Raw audio download whenever you need the source.
Security
Security by architecture, not by promise.
VoxSpace is designed so that sensitive voice data stays inside boundaries you already operate — your servers, your keys, your identity provider.
Self-hosted by design
Run VoxSpace on your own infrastructure — on-premises or in your cloud — with fully offline, air-gapped deployment when the network policy demands it.
Your models, your endpoints
Connect your own ASR and LLM services. No third-party cloud is required anywhere in the data path — transcription and summarization happen on infrastructure you choose.
Encrypted in transit and at rest
TLS everywhere. Recordings, transcripts, and summaries are encrypted at rest, so storage operators cannot read them directly from the database or object storage.
Enterprise identity
SAML SSO and SCIM user provisioning plug into your existing directory, alongside local user management and site-wide announcements.
Auditable
Audit logs record session lifecycle events — created, viewed, moved, exported, deleted — so compliance reviews have something concrete to read.
iOS App
VoxSpace for iPhone.
A native SwiftUI client for iOS 18 and later — iPhone only. It binds to your server, records anywhere, and puts your library in your pocket. Nothing else.
Scan-to-bind enrollment
QR-code-only onboarding — there is no field to type a URL. The server profile is verified against the scanned payload before the device ever binds.
Hardware-backed device identity
Secure Enclave P-256 keys (with a software fallback), Keychain-resident and never exported. Every request is signed with a short-lived DPoP proof.
Record anywhere
Microphone capture keeps working when the network doesn’t. Background uploads resume automatically on reconnect — no babysitting, no lost audio.
Your library on the go
Browse sessions on your server, read transcripts and structured AI summaries, and play back bounded, validated media.
Thoughtful details
Live Activities while you record, a share extension for importing audio from other apps, and automatic cache hygiene.
Privacy by default
No third-party SDKs, no analytics, no tracking. The camera is used only for QR scanning. Read the Privacy Policy.
Coming soon on theApp Store
The iOS app is not yet listed. This badge links to the App Store once it ships.
Self-Host
Deploy in minutes.
The full stack — web app, API, database, and object storage — ships as Docker Compose services. Bring a host, point your models at it, and it’s yours.
- Copy the example environment and fill in your values.
- Build the images with the staging Compose file.
- Run the database migrations, then start the stack detached.
Note: browser microphone capture requires HTTPS (or
localhost). For remote deployment, point a domain at your host and
terminate TLS before opening it to users.
Your keys. Your models. Your data.
$ cp .env.staging.example .env.staging
# edit .env.staging with your values
$ docker compose --env-file .env.staging -f docker-compose.staging.yml build
$ npm run docker:staging:migrate
$ docker compose --env-file .env.staging -f docker-compose.staging.yml up -d