Launching > Launch on a schedule
Launch via a schedule
How do I launch processes on a recurring schedule?
Section titled “How do I launch processes on a recurring schedule?”You can auto-launch processes on a schedule - once at a set time, or daily, weekly, monthly, or at any interval you need. Tallyfy can run the schedule itself, as a scheduled job you set up through the API or by asking an AI assistant. You can also use a middleware platform like Zapier, Make.com, or Power Automate, or write your own script that calls the Tallyfy API on a cron schedule.
What you’ll need
Section titled “What you’ll need”- A Tallyfy account with process launch permission (Admin, template owner, or granted launch access)
- A published template ready to launch
- One of these: an AI assistant connected to Tallyfy, access to the Tallyfy API, a middleware account (Zapier, Make.com, Power Automate), or a server where you can run scheduled scripts
Option 1 - a scheduled job that Tallyfy runs
Section titled “Option 1 - a scheduled job that Tallyfy runs”A scheduled job launches a process from a template on its own, once at a set time or on a repeat (hourly, daily, weekly, monthly or yearly). Tallyfy runs it, so you don’t need another tool or a server of your own. The web app doesn’t have a screen for scheduled jobs yet, so you set one up in one of two ways:
- Ask an AI assistant connected to Tallyfy’s MCP server, such as Claude. For example: “Launch the Equipment Safety Check template every Monday at 8 AM.”
- Call the Tallyfy API. The calls are listed under For developers below.
Times are read in your organization’s time zone.
Who a scheduled job runs as
Section titled “Who a scheduled job runs as”Every scheduled job has an owner, and it runs as that person. The owner starts out as the person who created the job, so each process it launches is started by them.
Before each run, Tallyfy checks that the owner is still an active member who can launch the template. If they aren’t, that run fails, and the job’s run history says why. A one-time job that fails is paused.
Changing a job’s owner
Section titled “Changing a job’s owner”An org admin can change who a scheduled job runs as, by asking their AI assistant or through the API. The new owner must be an active member who can launch the template, and Tallyfy Bot can’t be an owner. If the person you pick doesn’t qualify, Tallyfy refuses the change and the job keeps its current owner. Tallyfy also refuses any owner change while the job’s template is archived or deleted. Restore the template first. Through the API you can point the job at another template instead, but only while its current owner is still an active member who can launch it. If the owner has left too, restore the template, or delete the job and create a new one.
When the owner leaves your organization
Section titled “When the owner leaves your organization”When an admin deactivates someone and hands their work to another person on the Automatic conflict resolution page, that person takes over their scheduled jobs too. Only the jobs the new person can run move to them. Any others stay with the person who left, and fail.
If nobody takes over their work, their scheduled jobs stay with them and fail. So change the owner of each of their jobs before you deactivate them. See removing a member.
Option 2 - middleware platforms
Section titled “Option 2 - middleware platforms”No coding required.
- Connect your middleware platform to Tallyfy.
- Create a new workflow with a Schedule trigger.
- Set the frequency (time, day, interval).
- Add a Tallyfy Launch Process action pointing to your template.
- Set the process name format and fill any required kick-off form fields.
- Test and activate.
Example - daily equipment check:
Trigger: Schedule (Every day at 8:00 AM)Action: Launch Tallyfy processTemplate: Equipment Safety CheckProcess name: Equipment Check - {date}Option 3 - API with a cron job
Section titled “Option 3 - API with a cron job”Write a server-side script and schedule it with cron (Linux/Mac) or Task Scheduler (Windows).
- Write a script that calls
POST /organizations/{org}/runswith yourchecklist_id. - Schedule it using cron or another task scheduler.
- Add logging and error handling.
- Run the script once by hand, and check that the process appears in Tallyfy before the schedule takes over.
Example - Node.js script launching every Monday at 9 AM:
const cron = require('node-cron');const axios = require('axios');
cron.schedule('0 9 * * 1', async () => { try { await axios.post('https://go.tallyfy.com/api/organizations/YOUR_ORG_ID/runs', { checklist_id: 'your-template-timeline-id', name: `Weekly Report - ${new Date().toLocaleDateString()}` }, { headers: { 'Authorization': 'Bearer YOUR_API_TOKEN', 'X-Tallyfy-Client': 'APIClient', 'Content-Type': 'application/json' } }); console.log('Process launched'); } catch (error) { console.error('Launch failed:', error); }});Option 4 - trigger from existing systems
Section titled “Option 4 - trigger from existing systems”If you’re already using an ERP, HRIS, or another system with scheduling, you can have it call the Tallyfy API directly or go through middleware.
- Pick the system that already knows when the work is due, such as your HRIS for a new hire’s start date or your ERP for month-end close.
- Set up an integration from that system to Tallyfy.
- Map data from the source system to Tallyfy form fields.
- Let it run once, and check that the process launches with the right form field values.
Common use cases
Section titled “Common use cases”| Process type | Frequency | Example |
|---|---|---|
| Compliance audits | Monthly | An inspection checklist on the 1st of each month |
| Financial reporting | Weekly | Expense approvals at 4 PM every Friday |
| Team updates | Daily | A standup checklist at 9 AM, Monday to Friday |
| Maintenance checks | Quarterly | A facility inspection every three months, from the first day of a quarter |
For developers
Section titled “For developers”(Skip this unless you’re setting up the technical side.)
Scheduled jobs use the API’s recurring-jobs endpoints. Send the same headers as on any other call: Authorization: Bearer, X-Tallyfy-Client: APIClient and Accept: application/json.
GET /organizations/{org}/recurring-job-typeslists the kinds of job. Use the id of thelaunch-processtype.POST /organizations/{org}/recurring-jobscreates a job. The body needs the job’s ownname(up to 50 characters),recurring_job_type_id,start_date,scheduleanddata.end_dateandmax_occurrencesare optional, and stop a repeating job.start_dateis the day the schedule is counted from, today or later. The first run isn’t always on that day. To see when it is, read the job back withGET /organizations/{org}/recurring-jobs/{id}and check itsnext_occurrence.scheduleneedsrecurring_type(hourly,daily,weekly,monthlyoryearly) andstart_time(HH:MM), and takesinterval. A weekly schedule also needsdays_of_week(1 for Monday to 7 for Sunday). A monthly one needsday_of_month, orweek_of_monthwith one day indays_of_week.datatakes the template’schecklist_id, thenameeach launched process gets, and optionalprerunvalues for the kick-off form.GET,PUTandDELETE /organizations/{org}/recurring-jobs/{id}read, change and delete a job.GET /organizations/{org}/recurring-jobs/{id}/historyreturns its run history.PUT /organizations/{org}/recurring-jobs/{id}/ownerwith the body{"user_id": 123}changes the owner. Only an org admin can call it. It answers422and changes nothing when the new owner isn’t an active member who can launch the template, or is the organization’s bot. It also answers422when the job’s template was archived or deleted, with the message a failed run records: “The template this job launches was archived or deleted. Restore it, or pick another template.” If the job’s owner has also left the organization, the message says instead: “The template this job launches was archived or deleted, and the job’s owner can no longer run it. Restore the template, or delete this job and create a new one.”
start_date and start_time are read in the organization’s time zone. The owner is kept in the job’s data.created_by, and updating a job with PUT /organizations/{org}/recurring-jobs/{id} never changes it, even when an owner change lands while the update is running. Each job the API returns also carries an owner object with the owner’s id, full_name and email.
A one-time job is a daily schedule with max_occurrences set to 1, and its date and time must still be ahead. max_occurrences goes at the top level of the body, beside start_date. Inside schedule it’s ignored, and the job runs every day. This body launches one process at 9 AM on December 31:
{ "name": "Year-end audit", "recurring_job_type_id": 12, "start_date": "2026-12-31", "max_occurrences": 1, "schedule": { "recurring_type": "daily", "interval": 1, "start_time": "09:00" }, "data": { "checklist_id": "YOUR_TEMPLATE_ID", "name": "Year-end audit 2026" }}Replace 12 with the id of the launch-process type from GET /organizations/{org}/recurring-job-types.
Through the MCP server, schedule_process_launch creates a job and change_scheduled_job_owner changes its owner.
Related articles
Section titled “Related articles”Zapier > Automate monthly process launch with Zapier