Pricing that scales with the work
Every limit below is enforced by the platform, not by the honour system.
Free
Connect one system and see an AI employee work.
$0 $0
per month per month, billed annually
- Agents 1
- Connections 1
- Agent executions 100
Starter
Put agents to work across a handful of systems.
$49 $40.83
per month per month, billed annually
- Agents 3
- Connections 5
- Agent executions 5,000
Pro
Multi-channel agents with human approval.
$199 $165.83
per month per month, billed annually
- Agents 10
- Connections 20
- Agent executions 50,000
Business
Scale agents across the organisation.
$799 $665.83
per month per month, billed annually
- Agents 50
- Connections Unlimited
- Agent executions 250,000
Compare every plan
| Feature | Free | Starter | Pro | Business |
|---|---|---|---|---|
| Agents Maximum number of agents in the workspace. | 1 | 3 | 10 | 50 |
| Connections Maximum connected software integrations. | 1 | 5 | 20 | Unlimited |
| Team members Maximum members in the workspace. | 2 | 5 | 25 | 200 |
| Agent executions Agent executions per billing period. | 100 | 5,000 | 50,000 | 250,000 |
| AI messages AI model requests per billing period. | 1,000 | 25,000 | 250,000 | 1,000,000 |
| AI tokens Model tokens per billing period — input, output and cached input together. | 250,000 | 5,000,000 | 50,000,000 | 250,000,000 |
| Change history How long the record of what agents changed is kept. | 7 | 30 | 60 | 90 |
| Web chat Talk to agents in the dashboard. | ✓ | ✓ | ✓ | ✓ |
| Telegram Operate agents from Telegram. | — | ✓ | ✓ | ✓ |
| WhatsApp Operate agents from WhatsApp. | — | — | ✓ | ✓ |
| Slack Operate agents from Slack. | — | — | ✓ | ✓ |
| Approval workflows Human approval for high-risk actions. | — | — | ✓ | ✓ |
| Advanced permissions Per-member permission overrides. | — | — | ✓ | ✓ |
| Advanced reports Saved templates, scheduling and exports. | — | ✓ | ✓ | ✓ |
| API access Programmatic access with API keys. | — | ✓ | ✓ | ✓ |
| Webhooks Outbound event delivery. | — | ✓ | ✓ | ✓ |
| Advanced audit logs Extended retention, export and chain verification. | — | — | — | ✓ |
| SSO SAML/OIDC single sign-on. | — | — | — | ✓ |
| Priority execution Higher-priority queue placement. | — | — | — | ✓ |
| SLA Contractual uptime and support response. | — | — | — | — |
Prices and limits are read from the database at runtime — what you see here is exactly what the platform enforces.
The number worth reading twice
Most limits on this page cap how much you can build. The AI token allowance is different: when it runs out, anything that needs a model stops until the period resets or you move up a plan.
So it is shown as you spend it, the chat screen warns before the ceiling rather than after, and whoever owns the workspace gets one email when it is reached. Nothing is deleted, and everything that does not need a model keeps working.
usage
Upgrade mid-period and you pay the difference, not the whole price — the part of the month you already paid for is credited.