Memory and skills, how MyAgens learns your workflows
Two things make MyAgens improve over time instead of starting cold every session: a tiered memory, and a growing library of skills.
Memory that stays relevant
MyAgens keeps facts in tiers so the right ones are always at hand without drowning the agent in noise:
- Hot facts are injected every turn.
- Warm facts are recalled when they are relevant.
- Cold facts sit in the panel, out of the way.
From the panel you can promote or demote any fact by hand; left alone, facts cool on their own, dropping from hot to warm after a week or so unused and from warm to cold after about a month. Agents write memories as they work and recall them automatically. Optional local semantic recall, through Ollama or LM Studio, ranks memories by meaning rather than keywords.
Skills that build up
When an agent works out how to do something, it can save it as a skill, a reusable procedure it pulls up the next time a similar request comes in. The library grows as you use it, unused skills archive themselves, and you can export or import skills as portable files to move them between machines.
Portable
Both memory and skills export to plain files, so your fleet's knowledge is yours to back up and carry.
Get MyAgens
Self-hosted and open source. Install it on macOS, Linux, or Windows, then command your fleet of AI agents from Telegram or Slack.
More in Using MyAgens
