This is the most common point of confusion — and it matters.

Scale unlocks governance features within the Midfleet platform. You get the policy engine, semantic drift detection, the enforcement ladder (warn → guide → cooldown → block), and audit logs. You're running it yourself, on our cloud infrastructure, with standard support. It's the capability without the compliance wrapper.

Governance (Enterprise) is a separate plan that adds the trust and security layer that large organisations require before they can deploy software internally. That means: SSO/SAML + SCIM provisioning for IT-controlled access, self-hosted deployment so your data never leaves your infrastructure, bring-your-own database, custom data retention periods for compliance (SOC2, HIPAA-adjacent), multi-org isolation, SLA-backed support with a named contact, and procurement tooling (PO support, custom invoicing).

Scale is the right choice if your CTO or engineering lead wants to govern agents without bureaucracy. Governance is the right choice if your security, legal, or procurement team needs to sign off before deployment.

Scale gives you all the operational governance features — the things that let you control and audit what your agents do in real time. Policy engine, drift detection, enforcement, audit logs, e-Discovery, multi-workspace.

What it doesn't include is the infrastructure and compliance layer: no SSO/SAML, no self-hosted option, no custom data retention, no SLA contract, no multi-org isolation. Those are Governance (Enterprise) exclusives — they exist to satisfy procurement and security requirements, not just operational ones.

Coordination events are: handoff() fired, claim() created, blocker() surfaced, and context payloads transferred between agents.

heartbeat() and register() are always free — they never count toward your quota, on any plan. We don't penalise you for keeping agents alive or registering new ones.

You'll get an alert at 80% usage — never a surprise. On paid plans, events beyond your quota are billed at your tier's overage rate (Starter: $0.002/event, Growth: $0.001/event, Scale: $0.0005/event). On Free, coordination pauses until your next monthly cycle resets — no data is lost, agents just stop firing new events.

Yes. Upgrades take effect immediately and are prorated for the remainder of your billing cycle. Downgrades take effect at the start of your next cycle — you keep your current plan's features until then.

Yes. We have a startup program for early-stage teams (pre-Series A) and reduced pricing for open-source projects. Email [email protected] with a brief description of what you're building.

All paid plans start with a 14-day free trial. If you specifically want to trial the Scale plan with governance features enabled, reach out and we'll arrange it.

Yes. Midfleet sits at the coordination and governance layer — above the model layer. It doesn't care whether your agents run on Claude, GPT-4o, Gemini, Mistral, or any other model. It coordinates the infrastructure, not the intelligence. You can even run a mixed fleet with different models on different agents.

Midfleet's coordination layer is built on five primitives your agents call directly:

register() — An agent announces its identity, capabilities, and current state to the hub. Always free.

heartbeat() — An agent signals it's still alive. Always free. Used for liveness monitoring, never billed.

claim() — An agent claims exclusive ownership of a resource or task, preventing collisions.

handoff() — An agent transfers a task to another agent with full context preserved.

blocker() — An agent surfaces a blocking issue it cannot resolve autonomously, escalating to the team.

The Concierge is a routing and coordination agent that sits between your other agents and the Midfleet hub. It understands the current state of the fleet — who's working on what, what's blocked, what capacity exists — and routes tasks to the right agent at the right time. Think of it as the traffic controller, not the worker.

Drift detection is a Governance feature (Scale+). When you define a policy for an agent — its allowed scope, permitted actions, resource boundaries — Midfleet continuously monitors whether that agent's behaviour matches the stated intent.

When semantic drift is detected (the agent is doing something outside its defined scope), the enforcement ladder triggers automatically: Warn → Guide → Cool down → Block. Each step is proportional. The agent isn't killed outright unless it has exhausted every correction opportunity.

Crucially, this happens mid-flight — before the agent has completed the wrong action — not after the fact as a log entry.

Project integrations (Growth+): GitHub, Jira, Linear, Notion. Notifications: Slack. Docker sandbox for isolated execution is also available on Growth and above. More integrations are on the roadmap — reach out if you have a specific request.

No. Midfleet sits on top of whatever agent framework or infrastructure you already use — LangChain, CrewAI, custom implementations, direct API calls. It's additive, not a replacement. Your agents call Midfleet's primitives; everything else stays the same.

On Free, Starter, Growth, and Scale plans, data is stored on Midfleet's cloud infrastructure. On the Governance (Enterprise) plan, you can self-host Midfleet entirely on your own infrastructure — data never leaves your environment. You can also bring your own database.

SSO/SAML and SCIM provisioning are Governance (Enterprise) features. They're not available on Scale or below. If your IT team requires centralised identity management as a condition of deployment, the Governance plan is the right path.

We're working toward SOC2 Type II certification. If compliance documentation is a procurement requirement, reach out — we can share our current security posture and timeline. The Governance plan includes custom data retention designed to support compliance requirements.

Free: Community support — Discord and GitHub Discussions.

Starter: Email support with standard response times.

Growth: Priority email support.

Scale: Priority email + Slack shared channel with the Midfleet team.

Governance (Enterprise): Named support contact, contractual SLA on response and uptime, dedicated onboarding.

Sign up free — no credit card required. You'll be up and running with your first agent registered in under 10 minutes. The free plan supports up to 5 agents and 5,000 events per month, with all five coordination primitives included.

Email [email protected] with a brief description of your team size, the agents you're running, and what you need. No pitch deck, no demo call required unless you want one. We'll get back to you with the right information.