Usage & cost insights
See messages, tokens and estimated spend per agent at a glance — manage your budget and spot anything unusual yourself.
OSP.net roadmap
Here’s what’s in motion to make your agent more capable — grouped by where it is in the pipeline. Upvote the features you want most, and tell us anything we’ve missed. Your votes genuinely shape what we build first.
Recently shipped? See the changelog for release-by-release notes.
See messages, tokens and estimated spend per agent at a glance — manage your budget and spot anything unusual yourself.
Brand the whole experience yourself — upload a logo, pick your color, point your own domain — without waiting on us.
A no-code studio to design a bespoke agent — its purpose, personality, knowledge and tools — just by describing it in plain English, then deploy.
For larger teams: single sign-on, a full audit log, granular governance over which tools each person can use, and priority support — so your whole organization can run on agents with confidence.
A built-in health check that catches setup problems and shows them right in your dashboard — so you know your agent is working before you rely on it.
Let your agent act on a schedule or a trigger — a morning inbox digest, a follow-up when a new lead arrives — instead of only replying when messaged.
Review everything your agent does, and require your approval before sensitive actions like sending, deleting or spending.
View, search, pin and correct what your agent remembers — so it stays accurate and gets better over time.
Drop your agent onto your own website as a chat widget, turning it into front-line support for your customers.
Run separate agents for support, sales and personal tasks under a single account — each with its own memory and channel.
Try a fully working agent on us before you connect your own model key — so you can see the value first, then make it yours.
Continuous health monitoring with automatic recovery, so your agent stays available without you having to babysit it.
Connect tools like Google, Slack and more by approving access once, instead of copying and pasting keys.
On supported models, your agent can split a big job across specialist helpers and bring the results back together — so it researches faster, handles longer documents, and double-checks its own work before it replies.
Bring your team into one workspace where everyone has their own personal agent, but you share the skills, team agents and knowledge your business has built — with one bill and you deciding who gets access to what.
Tell us what would make your agent more useful. We read every request — the popular ones move up the roadmap.