Guide
Starting November 2024, every pricing plan has a monthly cap for resource-intensive services like automations, integrations, and AI:
This ensures fair resource allocation and incentivizes effective setups. The limits are rather generous and target only the heaviest of users: at the moment of introduction, only about 5% were affected. From now on, we aim to keep this number as the target: as Fibery's costs per run/sync go down and customers' usage goes up, we'll adjust the numbers.
How are the limits calculated?
Every plan has a base number of runs, syncs, and hours, plus some bonus ones for each paid seat on paid plans. Fibery is a collaborative tool, after all, and as your company scales and you invite more people, we increase your monthly cap.
For example, as of this writing, a 50-person company on a Pro plan gets 10,000 + 50 × 1,000 = 60,000 automation runs a month. If they invite another 10 people, the cap goes up to 70,000 monthly runs.
You can always see the up-to-date numbers for each plan on our pricing page.
Pro trial doesn't have paid seats, so it's capped at the base number, e.g., 10,000 runs.
The quotas are per calendar month and are not tied to your billing cycle. At the start of each month, the counters go back to zero (applies to the annual plan as well).
Where do we check how many we have left?
For the % of automation runs and integration entity syncs left in the current calendar month, navigate to Help & Support → Workspace status at the bottom of the sidebar:
Other limits are either displayed in-place like for finding patterns:
Or are they not currently displayed anywhere as barely anyone reaches them, as is the case with some of our (too?) generous AI limits?
Do we get notified if we are getting close to a limit?
For automations and integrations, we send two warning notifications to all workspace Admins:
75% of the limit reached before the 20th day of the month;
90% of the limit reached before the 25th day of the month.
We also display a sidebar banner to Admins suggesting an upgrade to the next plan:
What happens if we reach a limit?
When you use all of your monthly quota for a given service, it stops working:
Integrations stop syncing data from external tools.
Automation Rules stop running, and Buttons fail when users click them.
The rest of the workspace remains fully operational.
The affected service will work again once you upgrade or when a new month starts and the quotas are reset.
How do we avoid reaching the limits?
Most workspaces (at least those on paid plans) shouldn't worry about the monthly limits since the caps only affect the heaviest of users. However, if you find yourself approaching a certain limit, there are things you can do to optimize your usage.
Integrations
Entity sync happens every time an entity is created, updated, or deleted in Fibery based on the change in the source data:
if the same entity is updated 3 times during one month, it counts as 3 entity syncs;
it doesn't matter if an entity changed a little (just one Field) or a lot (ten Fields) since the last update: in any case, it is still considered a single entity sync;
if an entity hasn't changed since the last update, it's not counted towards entity syncs;
imported (as opposed to synced) entities are not counted towards the limit.
To lower the number of entity syncs:
filter the data source by excluding old or irrelevant data;
sync less often;
if you use a custom connector via Create Custom Integrations, make sure you have a delta sync that avoids unnecessary data updates.
Adding a new field in the source system does not trigger a new full sync in Fibery by default.
If you want Fibery to capture and store data from that new field, you’ll need to:
Go to your integration configuration in Fibery.
Run a full sync so that all existing entities are updated with values for the new field.
Regular syncs will continue as usual for the fields you already selected.
Automations
To lower the number of automation runs:
if you use the same Button to update many entities at a time, select them all (e.g. in a Table View) and click the Button once: for 100 entities, it's going to be 2 automation runs (50 entities per batch) instead of 100 runs when clicking the Button 100 times for each entity;
make sure your script actions work effectively with an array of incoming entities, as we can batch up to 50 entities together for a single automation run.
If none of the optimizations save you enough syncs/runs or you don't think they are worth your time, upgrade to the next plan for much higher or even unlimited quotas.
FAQ's
What counts as an “AI automation run”?
An AI automation run is counted every time an AI step is executed in an automation, regardless of the model used, prompt size, or outcome.
Do different AI models count differently toward my quota?
No.
All AI runs are counted equally, no matter which model you use.
✅ gpt-4o-mini, gpt-4o, or gpt-5 — each counts as 1 run
🚫 There’s no weighting by token usage, cost, or model complexity
The quota tracks number of invocations, not token volume
So:
1 run with gpt-4o-mini = 1 run with gpt-5 = 1 AI automation run
How are runs counted in chained or multi-step automations?
Each AI call inside your workflow counts as a separate run — even if triggered by one button or automation chain.
Button → AI Call 1 → AI Call 2 → AI Call 3
➡️ 3 runs (one for each AI step)
➡️ 5 runs total
Each step that invokes the AI counts individually toward your monthly quota.
Should I combine multiple tasks into one prompt, or split them?
It depends on your goals — here’s how to decide:
Approach | Pros | Cons |
|---|
Combine multiple tasks into one prompt | ✅ Counts as 1 run (saves quota)
✅ Good for related or simple tasks | ⚠️ Harder to parse structured output
⚠️ Less precise for multi-part workflows |
Split into separate prompts | ✅ More control over structure and logic
✅ Easier data extraction per field | 🥲 Each AI call = 1 run (uses more quota) |