Topic Scout
Turns your industry signals and past hits into angles you can shoot today.
You make the calls; your employees do the work. Each role gets its own agent — it knows your business, follows a fixed process, and hands back results you can use. Everything runs on your own machine, and your data never leaves it.
Free software · macOS, Windows and Linux
Interface demo · agents defined by your knowledge and process
It is smart, but it does not know your customers, your product, or the wording you settled on last month.
Every agent is three things: a role, a process, and your knowledge base.
Turns your industry signals and past hits into angles you can shoot today.
Replies to customers under your return policy, and flags anything it is unsure about.
Turns call recordings, chat logs and spreadsheets into structured notes and to-dos.
These are examples. Your agents are built for your business.
If the step has never been done before, and the rules live only in the owner's head, we come into your business and build it with you.
See who does this step today, how long it takes, and what it is worth.
Turn your rules, documents and judgment into a knowledge base and workflow the agent can use.
Run it on your machine and connect it to the process you already have.
You get the agent and its documentation; adjustments as needed.
The software is free. You only pay for agents: general ones per pack, custom ones per project.
Prices and terms may change before launch. Early access users keep first-batch terms.
The app runs locally and does not connect to our servers. But agents call a model to do their work, so your machine needs to reach the model provider you choose.
For the Personal plan, yes. You enter your own model API key in settings, and requests go straight to the provider, not through us. Custom deployments can be configured to your requirements.
Not to our servers. Knowledge bases, history and output files stay on your computer. When an agent calls a model, the content is sent to the model provider you chose — that is part of how the model works.
macOS, Windows and Linux. The first release focuses on macOS and Windows.
General chat starts from zero every time, so you keep repeating the context. A Levamen employee is bound to your knowledge base and a fixed process, so the same task is done to your standard every time, in your format.
We are in early access and not public yet. Leave your contact and we will tell you first when the first batch opens.
Agents work on your computer, not on our servers.
Bring your API key and call the model provider directly.
Knowledge bases, history and output files live in your own folder.
Describe the step and what it costs you today. We will tell you whether it is worth doing.
Add a note with "Levamen" and your industry.