Ad
Learn more

Open Source Machine Learning Infrastructure Tools

A curated collection of the best open source infrastructure components specifically designed for AI/ML workloads..

Run open-source LLMs locally on your own machine
  • Stars

    185,701
  • Last commit

    11 hours ago
  • License

    MIT
Runs open-source language models locally with a simple setup, plus optional cloud access for larger models when local hardware isn't enough.

Open Source Alternative to:

Run LLMs locally with minimal setup, maximum hardware support
  • Stars

    133,181
  • Last commit

    10 hours ago
  • License

    MIT
C/C++ inference engine for large language models, supporting quantization, multi-GPU, Apple Silicon, and an OpenAI-compatible server across a wide range of hardware.

Open Source Alternative to:

High-throughput, memory-efficient serving engine for LLMs
  • Stars

    Stars growing over the last 30 days96,158
  • Last commit

    10 hours ago
  • License

    Apache-2.0
Inference and serving engine for large language models, built for speed and hardware efficiency with an OpenAI-compatible API and support for a wide range of open models.

Open Source Alternative to:

Run open-source AI models privately on your own device
  • Stars

    77,347
  • Last commit

    1 year ago
  • License

    MIT
GPT4All runs open-source language models locally on Windows, macOS, and Linux with no cloud dependency, keeping your data on your machine.

Open Source Alternative to:

Self-hosted AI runtime for text, voice, vision, and agents
  • Stars

    50,139
  • Last commit

    10 hours ago
  • License

    MIT
Run LLMs, speech, image generation, and autonomous agents on your own hardware with an OpenAI-compatible API and 60+ swappable backends.

Open Source Alternative to:

AI-powered research notebook with full privacy control
  • Stars

    Stars growing over the last 30 days41,375
  • Last commit

    1 day ago
  • License

    MIT
Combines AI summarization, podcast generation, and multi-format content ingestion for researchers and students who want to keep their data private.

Open Source Alternative to:

Jan

AI
Run AI models locally or connect to cloud, privately
  • Stars

    45,332
  • Last commit

    10 hours ago
  • License

    Apache-2.0
Jan runs open-source AI models on your own hardware or connects to cloud providers like OpenAI, Anthropic, and Google, keeping your conversations off third-party servers.

Open Source Alternative to:

Open source LLM engineering platform for AI-powered applications
  • Stars

    Stars growing over the last 30 days36,898
  • Last commit

    10 hours ago
  • License

    MIT
Langfuse provides tracing, evaluations, prompt management, and analytics to debug and improve LLM applications.

Open Source Alternative to:

Open-source LLM tracing & evaluation for AI optimization
  • Stars

    Stars growing over the last 30 days12,106
  • Last commit

    10 hours ago
  • License

    Elastic-2.0
Open-source platform for LLM tracing, evaluation, and optimization. Features automatic instrumentation, prompt playground, and real-time AI application monitoring.

Open Source Alternative to:

Monitor LLM performance with open-source observability
  • Stars

    7,516
  • Last commit

    16 hours ago
  • License

    Apache-2.0
Open-source observability platform for LLMs using OpenTelemetry. Monitor performance, track costs, and debug AI applications with just 2 lines of code.

Open Source Alternative to:

Build reliable AI apps with comprehensive observability
  • Stars

    6,254
  • Last commit

    6 days ago
  • License

    Apache-2.0
Open-source platform for logging, monitoring, and debugging LLM applications. Route, debug, and analyze AI apps with comprehensive observability tools.

Open Source Alternative to:

Build, automate, and improve AI agents with your team
  • Stars

    4,881
  • Last commit

    10 hours ago
  • License

    MIT
Open-source workspace for building and running AI agents: chat-driven development, scheduling, human-in-the-loop approvals, versioning, and tracing. Self-host or use the cloud.

Open Source Alternative to:

Full observability for AI agents in production
  • Stars

    4,742
  • Last commit

    15 hours ago
  • License

    MIT
Open-source platform for monitoring AI agents: captures traces, surfaces failure patterns, alerts on issues, and helps you verify fixes with automated evals.

Open Source Alternative to:

Simulation-based testing and evaluation for AI agents
  • Stars

    5,004
  • Last commit

    10 hours ago
  • License

    Apache-2.0
Tests AI agents through multi-turn simulations, LLM-based scoring, and production tracing so teams can ship reliable agents with confidence.

Open Source Alternative to:

AI-powered platform for engineering LLM products
  • Stars

    Stars growing over the last 30 days3,471
  • Last commit

    11 hours ago
  • License

    Apache-2.0
Laminar is an open-source platform that helps collect, understand, and utilize data for building high-quality LLM applications.

Open Source Alternative to:

Python, SQL, and AI in a collaborative data notebook
  • Stars

    3,028
  • Last commit

    1 day ago
  • License

    Apache-2.0
Deepnote is an open-source collaborative notebook for data analysts and scientists, combining Python, SQL, AI assistance, and live sharing in one cloud platform.

Open Source Alternative to:

Monitor, debug, and scale LLM applications with ease
  • Stars

    Stars growing over the last 30 days2,899
  • Last commit

    10 hours ago
  • License

    Apache-2.0
Open-source observability platform for GenAI and LLM applications. Real-time monitoring, distributed tracing, prompt management, and AI model evaluation built on OpenTelemetry.

Open Source Alternative to:

GPU orchestration across clouds, Kubernetes, and on-prem
  • Stars

    2,299
  • Last commit

    13 hours ago
  • License

    MPL-2.0
Unified control plane for AI compute that provisions and schedules GPU workloads across any cloud, Kubernetes cluster, or bare-metal server.

Open Source Alternative to:

Serverless GPU compute with sub-second cold starts
  • Stars

    1,842
  • Last commit

    10 hours ago
  • License

    AGPL-3.0
Run GPU inference, task queues, and sandboxes on serverless infrastructure with sub-second cold starts, autoscaling, and support for your own AWS, GCP, or bare metal.

Open Source Alternative to:

Experiment tracking platform for ML teams
  • Stars

    394
  • Last commit

    8 months ago
  • License

    Apache-2.0
Open source ML experiment tracking platform with parameter and gradient logging, media tracking, real-time alerts, and full Weights & Biases API compatibility.

Open Source Alternative to:

LLM observability: cost, latency, and traces in one place
  • Stars

    13
  • Last commit

    1 day ago
  • License

    MIT
Drop-in observability platform for OpenAI, Anthropic, and Gemini that logs every request, tracks costs, traces agent workflows, and flags anomalies and PII.

Open Source Alternative to:

Trace, evaluate, and protect every LLM call in one SDK
  • Stars

    12
  • Last commit

    11 hours ago
  • License

    AGPL-3.0
An AI gateway that sits on the request path to handle PII redaction, prompt injection blocking, tracing, evals, and cost tracking for OpenAI, Anthropic, and Gemini.

Open Source Alternative to: