Inference
MITOllama
Run large language models locally with a simple CLI and REST API.
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.
Builder-first curation
Useful by workflow
Less noise, more signal
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.
Inference
Proprietary (free)LM Studio
Desktop app for discovering, downloading, and running local LLMs.
Inference
MITLocalAI
Free, open source OpenAI-compatible API server for local inference.
Serving
Apache 2.0vLLM
High-throughput LLM serving with PagedAttention for production use.
Frameworks
MITLangChain
Framework for building LLM-powered applications and agent pipelines.
Frameworks
MITLlamaIndex
Data framework for connecting LLMs to external data sources and RAG.
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
Llama 3 70B
Meta's flagship open weights model. Strong reasoning, coding, and instruction following.
Chat
Mistral 7B
Punches well above its weight class. Excellent for local inference on consumer GPUs.
Chat
Qwen2 72B
Top-tier open weights model with strong multilingual and coding capabilities.
Code
DeepSeek Coder V2
Best-in-class open source coding model. Rivals GPT-4o on coding benchmarks.
Edge
Phi-3 Mini
3.8B parameter model that outperforms models 3x its size. Runs on CPU.
Chat
Gemma 2 27B
Google's open weights model family. Strong performance at the 27B scale.
Code
Code Llama 34B
Meta's code-specialized model. Strong at completion, infilling, and instruction following.
Chat
Falcon 40B
UAE Technology Innovation Institute open model. Strong multilingual capabilities.
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
Why 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.
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.