Pricing
Plans for governed AI agent workflows.
Pick the billing-account plan that matches your MCP-connected agents, GitHub and Postgres/database tools, runtime sessions and logs, approvals, monitor/enforce posture, and monthly plus daily quota controls shared across workspaces.
Free
Start$0/month
For individual builders validating governed MCP access with a small workspace.
- 1 workspace included
- 2 members per workspace
- 2 agents
- 1 protected repository across account
- 1 active database connection across account
- 3,000 runtime calls/month shared across workspaces
- 300 enforce calls/month shared across workspaces
- 300 total calls/day
- 30 enforce calls/day
- 7-day logs
- Monitor + enforce
- 1 limited custom policy set
- Community/basic support
Team
Most Popular$20/month
For teams rolling out AI coding agents with shared policies and audit visibility.
- 3 workspaces included
- 10 members per workspace
- 5 agents
- 10 protected repositories across account
- 3 active database connections across account
- 50,000 runtime calls/month shared across workspaces
- 50,000 enforce calls/month shared across workspaces
- 2,500 total calls/day
- 2,500 enforce calls/day
- 30-day logs
- Monitor + enforce
- 5 custom policy sets
- Basic support
Growth
Scale$99/month
For organizations standardizing runtime controls across agents, repos, and environments.
- 10 workspaces included
- 25 members per workspace
- 20 agents
- 50 protected repositories across account
- 10 active database connections across account
- 150,000 runtime calls/month shared across workspaces
- 150,000 enforce calls/month shared across workspaces
- 10,000 total calls/day
- 10,000 enforce calls/day
- 90-day logs
- Monitor + enforce
- 25 advanced policy sets
- Priority support
Enterprise
CustomCustom
For regulated teams that need identity controls, exports, custom retention, and deployment support.
- Custom workspaces
- Custom members
- Custom agents
- Custom protected repositories
- Custom active database connections
- Custom runtime calls
- Custom daily and billing-period enforce limits
- SSO/SAML/SCIM
- SIEM export
- Custom retention
- VPC/self-hosted later
- Dedicated support
Prices are monthly. Cancellations take effect at the end of the billing period. Taxes may apply later if configured.
How plans differ
| Capability | Free | Team | Growth | Enterprise |
|---|---|---|---|---|
| Owned workspaces | 1 | 3 | 10 | Custom |
| Members/workspace | 2 | 10 | 25 | Custom |
| Agents | 2 | 5 | 20 | Custom |
| Protected repositories | 1 | 10 | 50 | Custom |
| Active database connections | 1 | 3 | 10 | Custom |
| Runtime tool calls/month | 3,000 | 50,000 | 150,000 | Custom |
| Enforce calls/month | 300 | 50,000 | 150,000 | Custom |
| Tool calls/day | 300 | 2,500 | 10,000 | Custom |
| Enforce calls/day | 30/day | 2,500/day | 10,000/day | Custom |
| Logs | 7-day logs | 30-day logs | 90-day logs | Custom |
| Runtime sessions | Included | Included | Included | Included |
| Approvals | Included | Included | Included | Included |
| Runtime modes | Monitor + enforce | Monitor + enforce | Monitor + enforce | Monitor + enforce |
| Policies | 1 limited custom policy set | 5 custom policy sets | 25 advanced policy sets | Custom policies |
| Rules per custom set | 5 | 25 | 100 | Custom |
| Support | Community/basic support | Basic support | Priority support | Dedicated support |
| SSO/SAML/SCIM | Not included | Not included | Later | Custom |
| SIEM export | Not included | Not included | Later | Custom |
| VPC/self-hosted | Not included | Not included | Later | Custom |
Built around runtime governance
Monitor before enforcing
Observe tool calls and would-be policy decisions before blocking production workflows.
Enforce high-risk boundaries
Move selected agents, actions, or repositories into enforce mode when policies are ready.
Keep the audit trail
Retain structured decision logs that show who called which tool, what policy matched, and what happened.
FAQ
Frequently Asked Questions
Everything you need to know before you start.
A billing account owns the plan, subscription, members, and shared usage pool. Workspaces are project/security boundaries inside that account.
Free accounts include one workspace. Additional workspaces require upgrading the billing account.
Runtime and enforce limits are shared across all workspaces in the billing account.
Normal signup creates one free billing account with one workspace. Additional workspaces require upgrading the billing account or contacting support@z-gateway.com for an organization setup.
Runtime quota is enforced at the billing-account level, so usage from one workspace reduces the remaining quota available to the others. Enforce quota exhaustion does not silently fall back to monitor mode.
Every runtime call routed through z-gateway counts toward your billing account's total tool-call quota. GitHub and database MCP tools share the same total runtime quota, and monitor and enforce calls both count. Enforce-mode calls also count toward enforce-specific limits where your plan has them. Free accounts include 300 enforce calls per monthly usage period, capped at 30 enforce calls per day.
Paid-plan monthly limits reset with the billing account's subscription billing period, so the reset follows the renewal date and is not always the calendar month. Free accounts use the app's UTC monthly quota window when no paid subscription period exists.
Total-call limits and enforce-call limits are checked separately. When a daily enforce limit is reached, enforce actions are unavailable until the daily window resets. When the monthly enforce limit is reached, enforce actions are unavailable until the monthly usage period resets. z-gateway does not silently fall back to monitor mode; users can explicitly choose monitor mode if they accept that actions will be logged but not blocked.
Protected repositories are the repositories selected in your GitHub App installation that z-gateway governs. You can manage selected repositories from your GitHub App installation, and your plan controls how many selected repositories z-gateway can govern.
Monitor mode evaluates policy and logs decisions without blocking denied actions. Enforce mode blocks denied actions before connector execution. Enforce availability depends on your plan and current enforce quota, and the mode switcher shows enforce usage before you switch.
z-gateway includes Postgres/database MCP connector support for implemented database actions such as schema inspection, bounded reads, explain/risk assessment, and approval-gated write proposals. Database connections are active protected database targets across the billing account; disabled or archived connections do not count toward your plan limit. Database actions count against shared runtime usage. It is not a transparent Postgres wire proxy and does not intercept arbitrary Prisma or psql traffic.
Risky or policy-gated actions can create workspace-scoped approval requests in the dashboard. Approval cards are designed to avoid exposing raw SQL or secrets in collapsed views, and execution follows the existing approval flow before connector execution for gated database actions.
Runtime sessions group related tool calls into one agent execution timeline, which is useful when an agent touches both GitHub and database systems. Detailed evidence still lives in runtime logs.
Most teams are connected in under 5 minutes. You add z-gateway as your MCP server URL, paste your agent credentials, and you're live. Full setup guide is in the docs.
Yes. z-gateway is MCP-native and vendor-agnostic. It works with MCP-capable clients such as Claude Code, Cursor, OpenAI Codex, and custom agents, with CLI-assisted setup available from the docs where supported.
Enterprise plans are available for custom limits, identity controls, exports, and private deployment discussions. VPC/self-hosted support is listed as later/custom, not a self-serve feature today. Contact us at sales@z-gateway.com to discuss your setup.
Yes. Every plan supports multiple members with role-based access control. You can invite team members as owners, admins, developers, or security viewers — each with different permission levels. The Team plan and above include unlimited seat access.
Yes. z-gateway supports PostgreSQL via the database MCP connector. You can define policies that restrict which tables an agent can query, block INSERT, UPDATE, or DELETE operations entirely, and get a full audit log of every query attempt — so an agent can never run a destructive operation you haven't explicitly allowed.
We are actively working toward SOC 2 Type II certification. If you have a specific compliance requirement or timeline, email us at sales@z-gateway.com and we'll work with you directly.