Use cases

RAGatm Open answers questions from your own documents and cites its sources. Here are typical situations, with the expected benefits and the limits to keep in mind.

Small and medium-sized businesses

Find the information buried in business documentation

Procedures, supplier contracts, reports, technical files: an SME accumulates documents nobody re-reads. One document space per topic lets staff ask questions in plain language and get the exact passage back.

Typical documents

  • Quality procedures (DOCX, PDF)
  • Supplier contracts and amendments
  • Minutes and annual reports
  • Datasheets and manuals

Expected benefits

  • Source-backed answers instead of keyword search
  • Documents hosted on your own server
  • Spaces partitioned by team or topic
  • Controlled hardware cost: a CPU server is enough for the Compact profile

Points of attention

  • A generated answer must always be checked against the cited source
  • Scanned documents without a text layer need OCR, which is not enabled in the Compact profile
  • Compact throughput suits a few users, not intensive load
Non-profit organisations

Make archives useful and volunteers’ lives easier

Bylaws, minutes, internal procedures, historical archives: non-profits keep their memory in scattered documents. A shared space lets new volunteers find answers quickly.

Typical documents

  • Bylaws and internal rules
  • General meeting minutes (PDF, DOCX)
  • Internal procedures and volunteer guides
  • Tracking spreadsheets (XLSX)

Expected benefits

  • Free software, no per-user licence
  • Runs on a modest server without a GPU
  • Separate spaces for the board, committees or archives
  • Documented backup and restore

Points of attention

  • Someone must be able to install and maintain a Docker stack
  • Paper archives must first be scanned with a text layer
  • Not a full document management system: RAGatm Open queries documents, it does not replace a DMS
Freelancers, researchers and individuals

Query a personal collection of notes and publications

Theses, papers, reading notes, technical documentation: a personal collection becomes searchable, with the assurance that nothing leaves your machine.

Typical documents

  • Papers and publications (PDF)
  • Markdown or plain-text notes
  • Reports and dissertations (DOCX)
  • Data exports (CSV, XML)

Expected benefits

  • Everything runs locally, models included
  • Common formats supported: text PDF, Markdown, DOCX, XLSX, PPTX, XML, EML
  • No data sent to a third-party service in the qualified configuration

Points of attention

  • The Compact profile needs about 14 GB of RAM and 8 CPU threads: check your machine
  • Compact models (1.5 billion parameters) produce useful but reasoning-limited syntheses
  • Audio and image files are not processed in the Compact profile
Developers and IT managers

Deploy a controlled document AI, module by module

Three repositories, documented API contracts, hash-locked dependencies and a reference Docker Compose stack: RAGatm Open is built to be audited, deployed and extended.

Typical documents

  • Internal technical documentation
  • Runbooks and operating procedures
  • Specifications and API contracts
  • Decision records (ADRs)

Expected benefits

  • Modular architecture: application, model gateway, optional pool
  • Base images pinned by digest, locked dependencies, non-root images
  • Typed configuration registry and startup checks
  • Versioned gateway API contract

Points of attention

  • No official Docker image is published: images are built locally
  • The gateway API contract is still a candidate
  • Scaling beyond one concurrent generation is not qualified

Ready to try?

The Compact profile installs on an 8-thread, 14 GB RAM machine, with no GPU.

See installation