Tier 2 Operator Guide — NeuralMind

Last updated: 2026-07-22
Version: v1.7.0
Audience: Operators who just ran neuralmind onboarding or neuralmind wakeup .

What Tier 2 (Team) gives you

Tier 2 is the paid tier on top of the MIT core. The MIT core’s 12-50x token reduction is free forever, and the tier2 feature set — governance, audit, self-hosted — runs under the auto-issued free 1-seat license too. A paid Team license buys seats beyond one (5-50), priority support, and an annual invoice, not hidden features.

Feature	Free license (1 seat)	Team ($29/user/mo, 5-50 seats)
12-50x token reduction	Yes	Yes
Synapse learning	Yes	Yes
MCP server	Yes	Yes
Memory publish/inherit	Yes	Yes
Governance (publish scope, weight threshold)	Yes, at 1 seat	Yes
Audit log (SHA-256 hash-chained)	Yes, at 1 seat	Yes
Self-hosted deployment	Yes, at 1 seat	Yes, with deployment support
Seats beyond one	No	Yes (up to 50)
Priority support + annual invoice	No	Yes
License validation (Ed25519 + 30-day grace)	Auto-issued free license	Signed Team license

You are paying for seats and support, not for features — everything is evaluable free at 1 seat.

Quick Start (first 5 minutes)

Install + activate free tier:

pip install neuralmind
cd /path/to/your-project
neuralmind build .
neuralmind wakeup .          # auto-issues free license

Run the onboarding wizard:
```
neuralmind onboarding --quick   # skip prompts, defaults only
```
Or just neuralmind onboarding for interactive mode. Walks through license activation, governance defaults, admin emails, and seat audit.

Verify:

neuralmind team license status
neuralmind team seats list

Tier 2 Commands

License

neuralmind team license status       # current tier, seats, expiry, issued-to
neuralmind team license activate <file>   # activate with Ed25519-signed license

The signed licence file comes from the vendor. If you are the vendor, the issuing side — quoting, invoicing, neuralmind license issue, renewals — is the Billing Runbook.

Seats

neuralmind team seats list            # list all seats
neuralmind team seats add <email>     # invite a seat (idempotent, soft-delete aware)
neuralmind team seats remove <email>  # remove a seat (soft-delete)

Governance

neuralmind team governance status
neuralmind team governance set-scope <personal|shared|both> --admin <email>
neuralmind team governance set-weight-threshold <0.0-1.0> --admin <email>
neuralmind team governance set-governance-enabled <true|false> --admin <email>
neuralmind team governance list-shared [--project PATH] [--json]
neuralmind team governance remove-edge SOURCE TARGET [--project PATH] --admin <email>

What each setting does (enforced since v4.6.0; earlier versions recorded scope and threshold without enforcing them):

Setting	Effect on `neuralmind memory publish`
`set-scope personal`	Publishing is refused (exit 1): memory stays on each machine
`set-scope shared`	Only the team baseline (the `shared` namespace) is published
`set-scope both` (default)	Personal + shared memory is published
`set-weight-threshold 0.1` (default)	Synapse edges below the threshold are left out of the bundle (transitions aren’t weight-filtered)
`set-governance-enabled false`	The scope and threshold gates are off; events are still audited

Settings live in ~/.config/neuralmind/tier2.yaml (per user, not per repository). A config that exists but can’t be read (unreadable, malformed YAML, an invalid value) makes publish refuse rather than publish ungoverned memory.

remove-edge SOURCE TARGET stops sharing one association: it is deleted from the project’s shared memory and review queue, dropped from .neuralmind-team-memory.json and listed under that file’s retracted key. Commit the file: each teammate’s next session deletes the pair from their own shared memory, and no later publish re-adds it. list-shared shows what is in shared memory now, strongest first. Walkthrough: Govern what your team’s agents share.

Audit

neuralmind team audit list            # recent audit events
neuralmind team audit export --format csv|json --output <file>
neuralmind team audit verify          # verify SHA-256 hash chain integrity

Recorded: governance configuration changes, seat and license actions, and (since v4.6.0) every team-memory publish (including refused ones), import of a teammate’s bundle, review_approve / review_reject, and remove. The actor is NEURALMIND_ACTOR_EMAIL (or --admin for admin commands), else the repository’s git config user.email.

Self-hosted

neuralmind team self-hosted init      # initialize self-hosted data dir (0700 perms)
neuralmind team self-hosted status
neuralmind team self-hosted validate-license

Honest scope (what Tier 2 does NOT do yet)

Claim	Status	Honest framing
SSO/SAML/OIDC	Not shipped	PRD-scoped for future wave
Real-time cross-machine sync	Not shipped	Paid tier future
Customer self-serve dashboard	Not shipped	Decide later
License portal	Deferred	D12: defer until 3+ customers — the manual path is the Billing Runbook

Do not market Tier 2 as SSO-enabled or multi-machine-synced. Architecture is ready; features are not shipped.

Troubleshooting

Symptom	Fix
“License not found”	`neuralmind wakeup .` auto-issues free tier on first run
“Downgrade guard triggered”	Fresh `tier2.yaml` should have `tier: free` (v1.7.0+). If not, delete and re-`wakeup`
“Seat limit exceeded”	`neuralmind team seats remove <email>` to free a seat
“Audit log empty”	Events fire on governance, seat and license operations and on team-memory publish/import/review/removal — and only once a tier2 config exists (`neuralmind onboarding`)
“Self-hosted data dir permission denied”	`neuralmind team self-hosted init` sets 0700 permissions

When to skip Tier 2

Under 5K lines, free-tier, inline-only, prompt-caching already covering you
No compliance/audit requirements
Solo developer (1-seat free tier is sufficient)

The honest boundary is documented on the NeuralMind site at https://neuralmind.uk/effectiveness/ — free-tier users whose codebases fall below the sweet spot should verify their own reduction ratio before purchasing.

Wiki v1.0. Next review: after SSO/SAML ships.