Free
For two people who want to see whether it holds.
Seats pay for the people using it. All-In puts the model bill on the same invoice, as credits. Bring your own model has you pay your provider directly — same product, same agents, different bill.
For two people who want to see whether it holds.
All of it, free for 14 days.
All of it, free for 14 days.
For teams that need the controls to be provable, not just present.
Same agents, same skills. The only difference is where inference is billed.
You have negotiated rates, strict data-residency needs, or want to tune cost per workflow yourself. Credits stop being metered.
You would rather not manage provider accounts and want one predictable per-seat line to budget against.
Start with credits included to prove value fast, then add your own keys once usage patterns are clear — the product does not change, only who carries the model bill.
These values come from the same catalog the product reads when a workspace is created.
| Limit | Free | Bring your own model | All-In | Enterprise |
|---|---|---|---|---|
| Seats | 2 | unlimited | unlimited | unlimited |
| Connectors | 1 | unlimited | unlimited | unlimited |
| MCP servers | 2 | unlimited | unlimited | unlimited |
| Scheduled tasks per person | 1 | 10 | 10 | unlimited |
| Taught skills | 10 | unlimited | unlimited | unlimited |
| Library files | 50 | 2,000 | 2,000 | unlimited |
| History | 30 days | unlimited | unlimited | unlimited |
| Monthly credits | 1,000 | Not metered | 2,500 | unlimited |
| Models | efficient | full choice | full choice | full choice |
| Your own model keys | not included | included | not included | included |
| Google Chat | not included | included | included | included |
| MCP access control | not included | included | included | included |
| Skill proposals | not included | included | included | included |
Credits pay for model usage that runs through Sykik — mostly work that happens in the background, such as tasks scheduled overnight. The pool is shared across the workspace rather than per seat. We do not route everything to frontier models: each job goes to the model best suited to it, which keeps consumption down without cutting the result. Allotment and top-up pricing are not final yet.
We currently onboard a small number of partners directly and set up the first workflow together.