# llmware-ai/llmware

Unified framework for building enterprise RAG pipelines with small, specialized models

Repository: https://github.com/llmware-ai/llmware
Canonical: https://ross.abutalabs.com/products/llmware
Homepage: https://llmware-ai.github.io/llmware/
Language: Python
License: Apache-2.0
License Family: permissive
Topics: parsing, retrieval-augmented-generation, agents, generative-ai-tools, llamacpp, llm, small-specialized-models, onnx, openvino
Last push: 2026-05-17T13:53:37+00:00

## Health v2 (maintenance only)
Score: 76/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 82, release rhythm 67, longevity 76
- inputs: {"age_days": 1069, "days_push": 108, "days_rel": 141, "gap_med": 30.5, "n_releases_24m": 3}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 14850, forks 2943 (observed 2026-08-28T04:11:07.797097+00:00)

## What it is
llmware is a Python framework for building enterprise RAG pipelines and LLM-based applications using small, specialized open-source models. It combines a catalog of 300+ prepackaged models (GGUF, ONNX, OpenVINO) with integrated document parsing, knowledge base creation, and agent orchestration, optimized for local on-device deployment.

## Use cases
- build a RAG pipeline over enterprise documents
- run LLM inference locally on a laptop or AI PC
- parse PDFs and office documents into a searchable knowledge base
- orchestrate multi-step agent workflows with small models
- query private documents with a local LLM without sending data to the cloud
- fine-tune or use specialized small models for enterprise task automation

## When to choose
- you need private, on-premise or on-device LLM applications with no data leaving the machine
- you want an all-in-one framework covering parsing, retrieval, prompting, and agents
- you want to run quantized small models efficiently across Windows, Mac, and Linux with GPU/NPU support

## When to avoid
- you only need a thin wrapper around a single cloud LLM API
- you require massive-scale distributed training or serving of large foundation models
- you need a framework tied to a specific stack like LangChain or a managed vector database

## Facets
- artifact type: framework
- maturity: active
- function: rag, llm-inference, agent-framework, parser, prompt-engineering, machine-learning, chatbot
- domain: large-language-models, developer-tools, self-hosted
- platform: python, windows, self-hosted
- tags: small-language-models, gguf, onnx, openvino, document-parsing, enterprise-ai, local-inference, llamacpp, vector-search, retrieval-augmented-generation, ai-agents, natural-language-processing, macos, linux, gpu

## Member repositories
- llmware-ai/llmware (main) score 76

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:11:07.797097+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-29T17:07:07.738041+00:00, confidence not recorded.
  - readme: https://github.com/llmware-ai/llmware (fetched 2026-08-28T04:11:07.797097+00:00, sha 6bc7f6039067)
  - homepage: https://llmware-ai.github.io/llmware/ (fetched 2026-08-29T08:05:04.223490+00:00, sha b5c8034ccc94)
  - registry_pypi: https://pypi.org/pypi/llmware/json (fetched 2026-08-29T08:05:04.233266+00:00, sha e7dc861e4b48)
- Data as of 2026-08-30T08:39:29.467469+00:00.