osFoundry / Platform
Schedule agents 24/7
Agents that run on a cadence — daily briefs, hourly checks, weekly reports. They wake themselves up.
What it is
Schedule agents 24/7 lets you turn any osFoundry agent into a recurring job. Instead of opening the app and typing a prompt every morning, you define a cadence — every hour, every weekday at 8am, the first of the month — and the agent wakes itself up, does the work, and delivers the result. Daily briefs, hourly status checks, weekly reports, end-of-day summaries: anything you'd otherwise remember to run by hand becomes a standing routine.
Inside osFoundry it works like cron for agents, but the "job" is a full agent run rather than a shell command. Each scheduled agent carries its own instructions, its own model choice, and access to the same chat, apps, code, knowledge, and data as the rest of your workspace. Because osFoundry is bring-your-own-key, scheduled runs call the model you configured — local, cloud, or self-hosted — directly against your own API keys, with no reseller markup on the tokens a nightly job burns. And because osFoundry is a local-first desktop workspace, a schedule can run against data that stays on your device rather than shipping everything to someone else's server.
This matters most for the work that's valuable precisely because it's regular. A brief is only a brief if it lands before your day starts; a check is only a check if it runs while you're not watching. Scheduling removes the human trigger, so the cadence you actually wanted — hourly, daily, weekly — is the cadence you get, without a per-seat subscription pricing you out of running as many routines as you need.
Key capabilities
- Set a cadence — hourly, daily, weekly, or a specific time and day — and the agent runs itself without you opening the app
- Standing routines like morning briefs, hourly checks, end-of-day summaries, and weekly reports run on their own schedule
- Each scheduled agent uses your own API keys against the model you chose — local, cloud, or self-hosted — with no reseller markup on the tokens
- Scheduled runs reach the same chat, apps, code, knowledge, and data as the rest of your workspace
- Local-first desktop execution means a schedule can run against data that never leaves your device, with no per-seat fees on how many routines you run
Use cases
- Operations lead: an agent compiles an 8am daily brief from your data sources so the summary is waiting before the workday starts
- On-call engineer: an hourly check runs a defined agent against a system or feed and only surfaces what changed
- Team manager: a weekly report agent runs every Friday afternoon, assembling the recurring status roll-up without anyone kicking it off
Frequently asked questions
How do scheduled agents run if osFoundry is a desktop app?
They run as recurring routines tied to your workspace and the agent's own instructions. Because osFoundry is local-first, a scheduled run can execute against data on your device rather than a remote server, and it calls the model you configured with your own API keys.
What cadences can I set — is it just daily?
No. You define the cadence: hourly checks, daily briefs at a set time, weekly reports on a chosen day, or other recurring intervals. The point is that the agent wakes itself up on whatever schedule the task actually needs.
Does running agents around the clock cost extra per seat or mark up my usage?
osFoundry has no per-seat fees, so you're not paying more to run more routines. Scheduled runs use your own API keys directly, so you pay your provider's rate for the tokens a job uses with no reseller markup on top.