Run models locally, then add a private browser-based chat layer.
A practical guide to the open-source AI stack
Find serious open-source AI tools and model families worth building with.
OpenSourcesAI helps developers, founders, and AI builders scan the open ecosystem faster, compare tools by workflow, and assemble a stack for local inference, coding, retrieval, orchestration, and deployment.
Updated for 2026
Builder-first curation
Sources recorded for review
Start with a stack
Pick a practical path before choosing every tool.
Open-source AI gets easier when you choose by workflow: local chat, coding help, private documents, production serving, or evaluation.
Use embeddings, vector search, and reranking for source-backed answers.
Test coding models against real repository tasks before standardizing.
Connect open models to home automation and private assistant workflows.
Popular local AI paths
Move from exploration to useful workflows.
These paths connect the directory to practical implementation choices.
DirectoryModels
Open model families by category, license, and hardware needs.
DirectoryToolsRuntimes, serving layers, frameworks, and workflow tools.
DirectoryIntegrationsApps, platforms, and services that connect AI models to real workflows.
ResearchCompareSide-by-side trade-offs for local AI choices.
How-toGuidesPractical setup guides for local and private AI workflows.
ShortlistsBestCurated picks for tools, hardware, and open AI stacks.
CommunityForumDiscuss model releases, local inference, and what you built.
DirectoryResourcesExternal Reddit groups, forums, GitHub discussions, and developer spaces for open-source AI builders.
Featured tools
Infrastructure and workflows developers actually use.
Start with tools that show up repeatedly in local AI setups, coding workflows, retrieval systems, orchestration experiments, and model routing stacks.
Local runner
MITOllama
Run open models locally with a simple CLI, model library, desktop app, and local API.
Local runner
Proprietary free tierLM Studio
Desktop app for discovering, downloading, chatting with, and serving local LLMs.
Local runner
Apache 2.0Jan
Open-source desktop app for running local AI models with a friendly ChatGPT-like interface.
Local runner
MITGPT4All
Local AI desktop and SDK project for running open models on consumer machines.
Local runner
Apache 2.0llamafile
Mozilla-backed project for packaging LLMs into portable executable files.
Local runner
MITllama.cpp
Core C/C++ inference project behind many local GGUF model workflows.
Chat workspace
BSD-3-ClauseOpen WebUI
Self-hosted AI platform and chat interface for Ollama and other model backends.
Chat workspace
MITAnythingLLM
Private AI workspace for documents, agents, and local or hosted model providers.
Chat workspace
MITLibreChat
Open-source multi-provider chat platform with a familiar assistant interface.
Chat workspace
ProprietaryMsty
Desktop AI workspace for local and cloud models with a user-friendly interface.
Chat workspace
Apache 2.0LobeChat
Open-source modern chat UI for multiple model providers and assistant workflows.
Coding assistant
Apache 2.0Continue
Open-source AI coding assistant for VS Code and JetBrains with configurable models and context.
Top AI models
Open model families shaping 2026 workflows.
Track the model ecosystems influencing local inference, coding, multilingual applications, agentic systems, and self-hosted product work.
Chat
Qwen3 235B A22B
Flagship open-weight Qwen3 MoE model often chosen for serious reasoning, coding, multilingual work, and agent experiments.
Reasoning
DeepSeek R1
Open model family commonly used as a reference point for reasoning-heavy open-weight workflows.
Code
DeepSeek Coder V2
Coding-specialized open model family worth testing for completion, refactoring, and coding assistant workflows.
Agents
Kimi K2
Large open-weight MoE model from Moonshot AI often discussed for agentic coding and tool-use workflows.
Agents
GLM-4.5
Open-weight MoE model family from Z.ai designed around agent, coding, and reasoning workflows.
Reasoning
MiniMax M1
Open-weight hybrid-attention reasoning model worth testing for long-context and agent-oriented workloads.
Chat
Llama 3 70B
Meta open-weight model family still commonly used as a baseline for local AI stacks and app prototypes.
Chat
Gemma 3 27B
Open-weight Gemma-family model worth testing where Google ecosystem support and local deployment matter.
Categories
Browse the stack by workflow, not just by name.
Use categories to move quickly from exploration to implementation, whether you are testing local models, wiring up RAG, evaluating toolchains, or planning deployment.
Local LLMsCoding AssistantsVector DatabasesAgent FrameworksModel ServingChat InterfacesEvaluationFine-tuningRAGInferenceAI GatewaysOpen Models
Latest guides
Practical guides for builders.
Start with hardware, model choice, RAG, evaluation, and agent safety.
Guide
Best Local LLM Setup for Windows in 2026
A practical Windows stack for running local models, chat UIs, coding assistants, and RAG experiments.
Read guideGuide
Ollama vs LM Studio vs Jan: Which Local Runner to Choose
A three-way guide to choosing a local model runner for CLI, desktop, and open-source workflows.
Read guideGuide
How to Build a Private AI Chatbot with Local Models
Plan a private chatbot stack using local models, retrieval, permissions, and self-hosted interfaces.
Read guideGuide
How to Choose a Model for Coding, RAG, Summarization, and Agents
A builder-first decision guide for choosing open models by task instead of hype.
Read guidePopular comparisons
Compare tools before you assemble the stack.
Use comparisons to choose the right layer for the job.
Ollama vs LM Studio
Ollama vs LM Studio
Choose between a scriptable local model runtime and a polished desktop app for testing open models.
Read comparisonOllama vs Open WebUI
Ollama vs Open WebUI
Understand the difference between a local model runtime and a self-hosted chat interface.
Read comparisonOpen WebUI vs AnythingLLM
Open WebUI vs AnythingLLM
Compare self-hosted chat UX against document-centered private AI workspaces.
Read comparisonContinue vs Aider
Continue vs Aider
Compare an IDE-native open coding assistant with a terminal-native AI pair programmer.
Read comparisonBest-of lists
Shortlists for common buying and build decisions.
Grouped recommendations for local LLM tools, coding assistants, vector databases, inference servers, and privacy-first stacks.
Best list
Best Open-Source AI Tools in 2026
A grouped shortlist of open-source AI tools for local models, RAG, coding, serving, and observability.
View listBest list
Best Local LLM Tools in 2026
A practical shortlist of tools for running, testing, and using open models locally.
View listBest list
Best Open-Weight Models for Coding
Coding model families worth testing for open coding assistants and software agents.
View listBest list
Best Models for Local RAG
Model choices for local retrieval, embeddings, reranking, and answer generation.
View listWhy open source AI
Better control, clearer trade-offs, and a stack you can shape.
Open-source AI is not just about cost. It changes how teams think about privacy, portability, iteration speed, and infrastructure choice.
Curated for builders
- Open-source or open-weight projects with clear utility.
- Useful for local AI, coding, retrieval, orchestration, or deployment.
- Strong documentation, ecosystem traction, or implementation value.
Compare by use case
- What problem the tool solves in a real stack.
- Whether it fits local development, internal tooling, or production.
- Why you would choose it over adjacent options.
Build your stack
- Combine runtimes, chat interfaces, vector search, and coding tools.
- Move faster from prototype to working internal tool.
- Keep more control over data, hosting choices, and platform risk.
Open-source vs open-weight
Check the license before you build on the model.
Open-weight models can be incredibly useful without being open source in the traditional software sense. Track the exact license, source, and usage limits before using any model in production.
Builder checklist
- Record the exact model checkpoint and license.
- Keep source URLs for every model and tool in the stack.
- Run your own task eval before trusting benchmark claims.
For builders & vendors
Get listed, sponsor a placement, or pitch a better workflow.
OpenSourcesAI is built for developers and founders evaluating open AI stacks. We accept relevant submissions, limited sponsored placements, and useful workflow suggestions.
Sponsor or submit your tool
Use this form to ask about sponsorships, featured placements, or editorial review.
For sponsorships, featured placements, tool submissions, or editorial review, use this form or email sponsors@opensourcesai.com. Form submissions are reviewed by OpenSourcesAI.