01Knowledge retrieval, inside your network
On-premises. For agents.
Your rules are enforced before anything is retrieved.
PremAgentic indexes the Markdown and text files your organization keeps, and gives your people and your AI assistants cited passages out of them. Who may read what is set in PremAgentic, folder by folder, and checked on every search.
$ prem groups add hr $ prem groups add engineering $ prem ingest ./docs/hr --principals group:hr --prefix hr $ prem search "band four compensation review" --as group:hr # hr reaches the salary bands $ prem search "band four compensation review" --as group:engineering # engineering does not
02In brief
Each folder carries an ordered list of allow and deny entries, set by an administrator. The first entry that names the caller decides, and no entry means no.
Superseded material is held back unless a caller asks for history.
A document that declares a machine author is held from agents until a person has reviewed it.
Past its stale-after date, a document is flagged for people and held from agents.
- Windows and Linux
- Stock PostgreSQL 14 or later
- MCP for agents
- An administration portal
An attorney asks in plain words and gets the passage that answers, with the file and the heading it came from, out of the Markdown and text files the firm keeps. A superseded draft is held back unless someone asks for history, so the answer is the current version, cited, and nobody hunts through folders for it.
Each person's assistant sees what that person may read and nothing more. The clinical folders are open to the clinical staff and closed to the front desk, because the folder rules are checked on every search, so one deployment serves the whole practice, and a procedure past its stale date is held from assistants and flagged for people.
The program's own files, on the program's own servers, with a local embedding model and no cloud service. Assistants read and never write. A folder cleared for one team stays closed to every other, and a folder marked never for hosted models is never served to an assistant whose model runs outside the network.
03Status
Coming soon.
PremAgentic is built by Agave Information Solutions in Scottsdale, Arizona, for organizations that want AI assistants working over their own files, on their own servers.