ENT For enterprise teams

Private dictation, deployed your way.Free under the MIT License. Services when you want them.

For teams that can't send voice to the cloud. Recognition and cleanup run on each Mac. You sign it, stage verified models, and roll it out with your own MDM and egress rules.

MIT LicenseNo activationNo accountApple silicon · macOS 15+

Fig. 01 — Menu-bar panel · 332 pt

01 How it's offered

Free software. Optional services.

The MIT License already covers your commercial use. Nothing to activate, nothing to renew. When you want help putting it on a fleet, LockedIn Labs does the work alongside your team.

The software · MIT License

Free for personal and commercial use.

The whole app, on every Mac you choose. No enterprise license to buy, and evaluating organizations can review the full source on request.

  • No trial. No word cap, no time limit.
  • No activation. No license key, no license server.
  • No account. Nobody signs in. No telemetry.
IncludedMIT

The services · LockedIn Labs

Optional, and agreed separately.

Pick the help you need. An agreement covers only the services and service levels you choose, never permission to use the software.

  • Deployment on your MDM, under your signature.
  • Terminology your users import on day one.
  • Maintenance, validation and support for the life of the rollout.
OptionalLockedIn Labs

02 Services

Five services. Take what you need.

Every engagement is work done with your tools and your people. The app itself stays the same MIT-licensed build on every Mac.

Optional services from LockedIn Labs and what each delivers
ServiceWhat we doYou receive
S·01DeploymentPackage the build and verified models for your Mac management workflow: signing and notarization with your Developer ID, root-owned model staging, a PPPC profile for Accessibility, and your egress rule.A deployment runbook, an artifact inventory with checksums, a rollback and removal procedure, and an administrator handoff.
S·02TerminologyTurn your names, acronyms and product terms into vocabulary rules, scoped to email, chat or code, without collecting anyone's dictation.A reviewed vocabulary CSV your users import in one step.
S·03MaintenanceTrack the pinned dependencies and model list, and rebuild when you decide to move.Each build with a fresh bill of materials and provenance record.
S·04ValidationTest delivery into the apps your teams use, offline behavior with egress blocked, password-field skipping, Copy only and recovery from History, against your pilot plan.Observed results your security and engineering teams can inspect.
S·05SupportHelp for your administrators and users through the rollout and after it.Service levels written into your agreement.

03 On every Mac

Your signature. Your MDM. Your rules.

Everything an administrator needs to package, approve, watch and remove the app. No internet needed once the speech model is installed; lock it down further with your own firewall rule.

Deployment sheet — managed Macscom.lockedinflow.app
Identity
com.lockedinflow.app · bundle id · preferences domain · Keychain service · log subsystem
Requirements
Apple silicon · macOS 15 or later
Entitlements
1 · audio input · hardened runtime
Signing
your Developer ID · notarized by you
Models (system)
/Library/Application Support/LockedIn Flow/Models
Ownership
root:wheel · directories 0755 · files 0644 · no ACLs · no symlinks
Model files
39 · pinned by revision, size and SHA-256 · verified before load
Model size
≈484 MB default · English Precision ≈614 MB more
Microphone
approved by each user on first recording
Accessibility
pre-approve with a PPPC profile for your signature · or Copy only
Network
none needed once models are installed · enforce with your egress rule
User data
~/Library/Application Support/LockedIn Flow · what's kept is encrypted
Settings
per-user preferences · no dictated text
Ships with
SBOM.cdx.json · BUILD-PROVENANCE.txt
Logs
subsystem == "com.lockedinflow.app" · never contain dictated text
Removal
app · data folder · preferences domain · Keychain item · tccutil reset

Signature

Built from a commit you reviewed.

Sign and notarize with your own Developer ID. Every build records its source commit and whether the tree was clean, so you deploy only what you checked.

Models

Staged once. Refused if altered.

Install the speech model in a root-owned location users cannot modify. If that folder fails its checks, the app refuses to load a model rather than fall back. It never downloads one.

Copy only

For fleets that withhold Accessibility.

Transcripts stay in LockedIn Flow until the user chooses Copy, and nothing is typed into other apps. No Accessibility access needed.

04 Pilot

Start with a measured pilot.

A credible evaluation starts with clear boundaries and acceptance criteria, and runs on synthetic content.

  1. 01

    Scope

    Pick the Mac models, macOS versions, destination apps, vocabulary and network policy in scope.

  2. 02

    Validate

    Test permissions, recording, delivery, offline operation, recovery, removal and retention with synthetic dictation.

  3. 03

    Decide

    Review the evidence, open risks, support ownership and rollout criteria before going beyond the pilot group.

Bring your requirements

What to send us.

  • Your managed-Mac baseline and the apps in scope.
  • Your security questionnaire.
  • The support boundary you want.
Keep it out of the formNever send credentials, private vocabulary, recordings or regulated data through the website. Share a non-sensitive outline; the details come later, through channels you choose.

05 Procurement

Straight answers for buyers.

Do we need an agreement to use it commercially?

No. The MIT License covers personal and commercial use. An agreement with LockedIn Labs covers only the services you choose.

Who signs and distributes the app?

You do. Organizations sign with their own Developer ID, notarize, and deploy through their own MDM. Request a build for your team to start.

Does it need network access?

No internet is needed once the speech model is installed. The app never downloads models, checks for updates, sends telemetry or calls a hosted service; its speech library's downloader is switched off before any model loads. Deny it outbound access with your firewall rule if you want that enforced.

What does it need on each Mac?

An Apple silicon Mac on macOS 15 or later, about 484 MB for the speech model, the microphone permission, and Accessibility to type at the cursor. Copy only mode works without Accessibility.

Can our security team review it?

Yes. Read the security architecture, and ask for the threat model, bill of materials and full source as part of your evaluation.

LF LockedIn Labs

Bring private dictation to your team.

Define a pilot around your apps, terminology and endpoint policies, with results your security and engineering teams can inspect.