Pay for the platform.
Not for the tokens.
Monthly plans, cancel any time. Bring your own model keys or subscription, so model usage is never marked up.
For founders putting their first agents to work.
- 1 company · 5 agents · 2 seats
- 2,000 runs / month
- 20 sandbox hours / month
- Tasks, triggers & cron
- Approvals, guardrails & budgets
- Memory, knowledge base & skills
- Claude Code & Codex on your subscription
- In-app, web push & e-mail notifications
For small teams running real work on agents.
- 3 companies · 25 agents · 10 seats
- 10,000 runs / month
- 100 sandbox hours / month
- Everything in Starter, plus:
- Agents that hire agents
- Git workspaces & pull requests
- Recorded browser flows
- Telegram, Slack & Discord
- MCP endpoint & access tokens
- Daily run-error analysis
For companies where agents are part of operations.
- 10 companies · unlimited agents · 25 seats
- 50,000 runs / month
- 400 sandbox hours / month
- Everything in Team, plus:
- 3-year audit log
- Priority support
- Onboarding session
Custom limits, a dedicated environment and contract.
- Custom companies, seats & runs
- Dedicated environment
- Custom DPA & security review
- Invoicing on account
- Named contact person
14-day free trial on every plan · No credit card to start · Model usage billed by your provider
Compare plans
| Feature | Starter | Team | Business |
|---|---|---|---|
| Company & agents | |||
| Companies | 1 | 3 | 10 |
| Agents | 5 | 25 | Unlimited |
| Seats (team members) | 2 | 10 | 25 |
| Org chart, managers & roles | |||
| Sub-agents | |||
| Agent versions | |||
| Agents that hire & manage agents | — | ||
| Work | |||
| Runs per month | 2,000 | 10,000 | 50,000 |
| Tasks, triggers & cron schedules | |||
| Company chat with attachments | |||
| Message agents during a run | |||
| Recorded browser flows | — | ||
| Coding agents | |||
| Sandbox hours per month | 20 | 100 | 400 |
| Claude Code & Codex on your own subscription | |||
| Automatic Claude ↔ Codex fallback | |||
| Git workspaces, tests & pull requests | — | ||
| Knowledge | |||
| Memory that learns after every run | |||
| Knowledge base & Obsidian vault | |||
| Skills (SKILL.md import/export) | |||
| Safety & control | |||
| Approvals with allow / ask / deny policies | |||
| Isolated sandboxes & network allowlist | |||
| Budgets & kill switch | |||
| Audit log retention | 30 days | 1 year | 3 years |
| Integrations | |||
| All model providers (bring your own key) | |||
| MCP servers per agent | |||
| GitHub & GitLab connections | |||
| In-app, web push & e-mail | |||
| Telegram (with approvals), Slack & Discord | — | ||
| MCP endpoint & personal access tokens | — | ||
| Insight | |||
| Token & cost tracking per run | |||
| Run log export | |||
| Daily run-error analysis | — | ||
| Support | |||
| Help center & support chat | |||
| E-mail support | |||
| Priority support & onboarding | — | — | |
| Data processing agreement (DPA) | |||
Questions
Are model costs included?
No. You connect your own API keys (Anthropic, OpenAI, xAI, OpenRouter, …) or your existing Claude or ChatGPT subscription, and pay the provider directly. Mandistry shows every token and every cent per run, and budgets stop runs before they overspend.
What counts as a run?
One piece of work by one agent: from picking up a task, chat message or trigger until it reports back. Sub-agents count as their own runs. Retries of the same run do not.
What is a sandbox hour?
Coding agents (Claude Code, Codex, git workspaces) work in an isolated container. You use sandbox time while such a container is running. Agents that only call tools and models don't use sandbox hours.
What happens when I reach a limit?
We notify you at 80% and 100%. New runs then wait until the next billing period or until you upgrade; nothing is billed automatically on top of your plan.
How does billing work?
Plans are billed monthly in advance and renew automatically. You get a VAT invoice for every payment. Upgrades take effect immediately (pro rata); downgrades from the next period.
Can I cancel any time?
Yes. Cancel in your account settings before the next billing date and your plan stops at the end of the current month. You can export your data until then.
Prices with or without VAT?
Prices are shown excluding 21% Dutch VAT (BTW); use the toggle to see them including VAT. EU businesses with a valid VAT number outside the Netherlands are invoiced with VAT reverse-charged.
Who can buy Mandistry?
Mandistry is sold to businesses (B2B). By subscribing you confirm that you act for a company or as a self-employed professional.
Is there a free trial?
Yes. Every plan is free for 14 days. You only add payment details when you decide to continue.
Still unsure? Talk to us or.