The Best Self-Hosted AI Stack for Small Teams in 2026
An opinionated reference architecture for self-hosted team AI in 2026: vLLM, Open WebUI, LiteLLM, Qdrant, and Langfuse on one GPU box, with honest alternatives at every layer.
Discover the best open-source and developer-focused AI tools. LLMs, agents, image generation, RAG, and more, curated and rated for indie makers and power users.
Curated
All-in-one desktop and Docker app for private LLM chat with your documents.
Feature-rich web UI for Stable Diffusion with extensive extension ecosystem.
Week of Sep 21 to Sep 27, 2026
One tool from every category, reshuffled every Monday.
Original 3D Gaussian Splatting implementation for real-time radiance field rendering.
Multi-agent framework that turns a one-line requirement into designs, documents, and working code.
Accurately lip-sync videos to any audio using a pre-trained model.
Open-source AI software developer agent formerly known as OpenDevin.
Open-source MLOps platform that orchestrates ML pipelines and AI agents across any backend.
Hugging Face library for state-of-the-art NLP models
Start here
Inference
Agents
Coding
Vector DBs
Documents
Speech
The live index
Official Apache-2.0 SandBase CLI and local MCP bridge that connects AI coding agents to 2,000+ AI models through one account. It supports Cursor, Claude Code, Codex, Windsurf, Gemini CLI, OpenCode, an
Meta's promptable model that recovers full-body 3D human meshes with hands and feet from one image.
Arcee AI's US-trained 400B sparse MoE family under Apache 2.0, with a dedicated reasoning variant.
29 tools
58 tools
19 tools
37 tools
24 tools
48 tools
26 tools
28 tools
17 tools
39 tools
13 tools
12 tools
18 tools
49 tools
66 tools
47 tools
34 tools
31 tools
26 tools
31 tools
30 tools
34 tools
52 tools
18 tools
31 tools
44 tools
7 tools
An opinionated reference architecture for self-hosted team AI in 2026: vLLM, Open WebUI, LiteLLM, Qdrant, and Langfuse on one GPU box, with honest alternatives at every layer.
Exact model, quant, and engine picks for every VRAM tier in 2026, from an 8GB laptop GPU to a dual-3090 tower, with honest notes on what each setup cannot do.
Wan 2.2, LTX-2, and HunyuanVideo 1.5 lead the open-weight video field in 2026. What actually runs on your GPU, and which model to pick for each job.
Help the community discover amazing tools by submitting your favorites. Every submission is reviewed before publishing.
Submit a Tool