# ogx-ai/ogx

Open GenAI Stack

Repository: https://github.com/ogx-ai/ogx
Canonical: https://ross.abutalabs.com/products/ogx
Homepage: https://ogx-ai.github.io/
Language: Python
License: MIT
License Family: permissive
Last push: 2026-08-26T20:47:39+00:00

## Health v2 (maintenance only)
Score: 86/100 (v2, computed 2026-09-03T02:39:23.370411+00:00)
- activity 99, release rhythm 86, longevity 57
- inputs: {"age_days": 799, "days_push": 7, "days_rel": 13, "gap_med": 7.0, "n_releases_24m": 81}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 8416, forks 1363 (observed 2026-08-28T04:10:21.064452+00:00)

## What it is
OGX (formerly Llama Stack) is an open-source, OpenAI-compatible agentic API server and Python library that bundles inference, vector stores, file storage, moderation, tool calling, and agent orchestration into one process. It works with any model provider (Ollama, vLLM, Bedrock, Azure, etc.) and can be deployed as an HTTP server or imported as a library.

## Use cases
- self-host an OpenAI-compatible API server for local models
- swap between LLM providers without changing application code
- build AI agents with tool calling and vector store search
- point Claude Code or Codex at alternative model backends
- run a drop-in replacement for the OpenAI API on my own infrastructure
- add moderation and file storage to an AI app without sidecar services

## When to choose
- you want an OpenAI-compatible server you fully control and can run anywhere
- you need inference, vector stores, moderation, and agent orchestration in a single process
- you want model-agnostic routing across providers like Ollama, vLLM, Bedrock, or Azure
- you prefer a Python library option instead of running a separate gateway

## When to avoid
- you only need a lightweight inference proxy or load balancer rather than a full stack
- you are locked into a single vendor's managed API and don't need portability
- you need a specialized high-throughput inference engine rather than an API layer over providers
- your stack is non-Python and you only want the library mode rather than the server

## Facets
- artifact type: framework
- maturity: active
- function: llm-inference, agent-framework, rag, api-framework, http-server, vector-database, chatbot, sdk
- domain: large-language-models, apis, self-hosted, developer-tools
- platform: python, self-hosted, cross-platform, cli
- tags: openai-compatible, llama-stack, api-server, pluggable-providers, model-agnostic, tool-calling, moderation, vector-stores, ai-agents, retrieval-augmented-generation, docker

## Member repositories
- ogx-ai/ogx (main) score 86

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:10:21.064452+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 2026-08-29T17:27:17.419118+00:00, confidence not recorded.
  - readme: https://github.com/ogx-ai/ogx (fetched 2026-08-28T04:10:21.064452+00:00, sha 910fa1fbd139)
  - homepage: https://ogx-ai.github.io/ (fetched 2026-08-29T08:27:10.240144+00:00, sha be002c45b9b7)
  - registry_pypi: https://pypi.org/pypi/ogx/json (fetched 2026-08-29T08:27:10.243208+00:00, sha c4902bfd43e1)
- Data as of 2026-08-30T08:39:29.467469+00:00.