Umans Code for Organizations

Documentation page covering how Umans Code works for teams and organizations, extending the individual user guide. Focuses on three core concepts: seats for human team members (flat-rate), service-account keys for automations (pay-per-token from a shared org wallet), and unified billing.

At a Glance: Three Core Concepts

1. Seats for human team members — flat-rate per person with generous built-in usage.
2. Service-account keys for automations — pay-per-token, drawing from a shared org wallet.
3. One invoice — both seat subscriptions and service-account usage settle together on the same organization billing.

Seats (for Humans)

Admins purchase seats from the Billing → Members interface and assign them to teammates. Each seat includes:

• "Generous usage built in, 4 parallel sessions (any 10-minute window)."
• Access to the same four models available to individual plans: umans-coder, umans-kimi-k2.7, umans-flash, and umans-glm-5.2.
• Teammates follow the standard Quick Start flow — running umans claude is sufficient.

Pricing Tiers

Sessions Pool Across the Organization

When a team exceeds a seat's 4-session limit, extra sessions cost $20 each. However, "unused capacity from other seats in the same org offsets overages at 50%": two unused session-slots cancel out one extra session. If spare capacity covers the overage, nothing is charged.

Service Accounts (for Automations)

Service-account keys operate without any human behind them. They consume from the organization wallet at per-token rates, meaning you pay only for what an automation actually uses.

Key restriction: "Token-based billing is available only to organizations, and only through service-account keys." Personal accounts remain on flat-rate Pro/Max subscriptions.

Common Use Cases

• Production alert triage & mitigation — A bot watches alerts from PagerDuty, Sentry, or Grafana, pulls repository context, and drafts a mitigation plan before waking a human. Cost: dollars of tokens per incident rather than a full seat.
• Automated code review on every PR — Runs a review pass using the same models engineers use in their IDE. Typically far cheaper than hosted alternatives.
• Scheduled maintenance — Migration passes, dependency bumps, flaky-test investigations — any task that could be handed to an agent running on a cron schedule.

The shared pattern: non-interactive, bounded token budget, measurable value per run.

Token Pricing (Service-Account Usage)

Per-model rates in USD per 1 million tokens. Each request debits exactly what the model consumed (input, output, cache reads, cache writes) with no per-call minimums.

Important Distinctions

• Seat usage remains flat-rate — none of the above token pricing applies to human seats.
• Wallets can be topped up manually or set to auto-refill from Billing → Wallet.
• Per-key usage breakdowns are available in the ledger view under Billing → Wallet.

Getting Started — Step by Step

1. Contact the team. Organizations are provisioned manually. Email contact@umans.ai with your team size and the first admin's email.
2. First admin accepts the invitation. The org is created and an invite link is sent to the named admin, who signs in to activate it.
3. Admin buys seats via Billing → Members → Get More Seats.
4. Invite teammates. They sign in and run umans claude normally — no extra setup required.
5. Create a service account for any automation via Billing → Wallet → Create service account. Store the key like any other secret and point your bot at https://api.code.umans.ai.

Support Channels

• Discord — community Q&A at discord.gg/Q5hdNrk7Rw
• Email — contact@umans.ai
• Dashboard — manage the org at app.umans.ai/billing

Footer / Site Navigation

• Tagline: "Umans · code. Built for serious agentic work."
• Links: User Guide, Terms, Privacy, Contact
• Social: Discord, X (Twitter), LinkedIn — all under the umans_ai handle