Plans & limits
What Free and Pro include, how runs are counted and reset, and the limits that apply on every plan: run length, loops, schedules and webhooks.
A plan is four numbers. Everything else in Coworkly is the same on every plan: workflows, schedules, triggers, the full integration library, MCP and the REST API.
What a plan sets
| Free | Pro | |
|---|---|---|
| People | 1 | 5 seats |
| Runs at a time | 1 | 10 |
| Runs a month | 500 | Unlimited |
| AI credits | 100 when you sign up | 1,000 every billing cycle |
- People. A Free workspace is yours alone and takes no invites. Pro adds seats; pending invites count toward the five. See workspaces.
- Runs at a time. When more runs start than the plan runs at once, the extra ones wait in line and start as slots free up. Nothing is dropped.
- AI credits. Credits pay for the in-canvas assistant and for AI steps that use Coworkly's models. When they run out, top up in Settings → billing on either plan. Top-ups do not expire.
Workflows are not limited on any plan: build, activate and schedule as many as you like.
How runs are counted
One run is one execution of a workflow, whatever started it: a webhook, a schedule, an email, a text, a form, the Run button, the API or MCP. A run with two steps and a run with two thousand both count as one.
- A workflow that calls another workflow starts a second run, and each one counts.
- A run that pauses for an approval or a wait and later resumes is still the same run.
- Runs count whether they finish, fail or are cancelled.
- Draft runs from the canvas count like any other.
The count covers the calendar month and resets at 00:00 UTC on the first (see the two cycles below). On Free, once the month's 500 are used the next run is refused with a message that says so; a webhook caller gets an error response. Runs already in progress finish. Pro has no monthly number.
The two cycles
Two different clocks run, and they are not the same day for most workspaces.
- Runs follow the calendar month. Free's 500 runs are counted from 00:00 UTC on the first of the month to the end of that month, whenever you signed up. The workflows page shows the count and the day it resets.
- Pro follows your subscription. Pro renews one month after the day you subscribed, on the same day each month. The 1,000 AI credits arrive with each renewal and are for that month: what is left when the next renewal comes does not carry over. Credits you bought as a top-up are separate and never expire; the monthly credits are spent first.
Limits that are the same on every plan
These protect the platform and every workspace on it. They are not part of pricing and do not change when you upgrade.
| Limit | Value | What happens past it |
|---|---|---|
| Steps in one workflow | 100 | The canvas tells you when a workflow is full and does not place another step; the assistant and MCP clients are told the limit up front and are refused past it. Editing or removing steps always works. Split a large workflow into smaller ones joined by a call workflow step. |
| Longest single run | 6 hours | The run is stopped. Time spent paused for an approval or a wait does not count; the clock restarts when the run resumes. |
| Items in one loop | 10,000 | The loop step fails before its first iteration. Filter or paginate upstream. |
| Fastest schedule | Every minute | A schedule cannot fire more often. |
| Workflows calling workflows | 3 levels deep | The call step fails. |
| Webhook deliveries | 600 a minute per workspace | Extra deliveries are refused; the sender can retry. |
| Uploaded file size | 25 MB per file by default | The upload is refused. The step can set a different size. |
| Emails sent by workflows | 100 a day per workspace | The send step fails for the rest of the day. |
| Texts sent by workflows | 50 a day per workspace | The send step fails for the rest of the day. See text messaging. |
| MCP requests | 120 a minute per token | The request is answered 429 with a Retry-After. See MCP. |
How long things are kept
Finished runs, with their step inputs, outputs and logs, are kept for 30 days and then removed. Runs that are still pending, running or waiting are kept until they finish. See retention.
Enterprise
An Enterprise deployment is your own isolated Coworkly instance, on its own database and infrastructure, with limits and concurrency set for you. Write to support@coworkly.ai.