# HazyResearch/minions

Big & Small LLMs working together

Repository: https://github.com/HazyResearch/minions
Canonical: https://ross.abutalabs.com/products/minions
Language: Python
License: MIT
License Family: permissive
Last push: 2026-03-12T05:22:40+00:00

## Health v2 (maintenance only)
Score: 52/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 71, release rhythm 35, longevity 41
- inputs: {"age_days": 576, "days_push": 174, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1350, forks 151 (observed 2026-08-28T04:04:28.059125+00:00)

## What it is
Minions is a communication protocol and Python framework that lets small on-device LLMs collaborate with frontier cloud LLMs to cut cloud inference costs while preserving quality. It includes demo apps, a CLI, a WebGPU app, and an end-to-end encrypted Secure Minions Chat variant.

## Use cases
- reduce cloud LLM API costs by processing long contexts locally
- run a local small model that orchestrates a frontier cloud model
- build privacy-preserving chat where sensitive data never leaves the device
- experiment with local-remote LLM collaboration protocols from the research paper
- estimate inference costs across local and cloud model combinations
- run a demo chat app combining Ollama and cloud providers like OpenAI

## When to choose
- you want to lower LLM costs by offloading long-context work to an on-device model
- privacy matters and you want sensitive context processed locally
- you want to reproduce or build on the Minions research protocol
- you have a capable local machine and access to at least one cloud LLM API

## When to avoid
- you have no local hardware able to run a small model
- you need a production-ready, battle-tested inference framework
- you only want a simple cloud API wrapper without local models
- you need Python 3.13 support

## Facets
- artifact type: framework
- maturity: active
- function: llm-inference, agent-framework, rag, chatbot, cli
- domain: large-language-models, artificial-intelligence, privacy, developer-tools
- platform: python, cli
- tags: on-device-llm, cloud-llm, cost-optimization, local-remote-collaboration, ollama, secure-chat, end-to-end-encryption, ai-agents, macos, linux, docker

## Member repositories
- HazyResearch/minions (main) score 52

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:04:28.059125+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 2026-08-30T04:42:15.059730+00:00, confidence not recorded.
  - readme: https://github.com/HazyResearch/minions (fetched 2026-08-28T04:04:28.059125+00:00, sha 650d4da71cd2)
- Data as of 2026-08-30T08:39:29.467469+00:00.