MNN
alibaba/MNNMNN: A blazing-fast, lightweight inference engine battle-tested by Alibaba, powering high-performance on-device LLMs and Edge AI.
Pushed yesterday.Apache License 2.0C++Developer tools16.5k
The Stacks
Wing of Developer tools
Libraries, CLIs and infrastructure for building with AI.
Free to explore · curated by OneWave AI
215 developer tools projects in this wing of the Stacks. Review status is shown for every entry.
Regular searchNames, topics, and keywords
AI searchDescribe a task, find the right tools
AI searches are sent to OpenAI.
Review is in progress. 1,407 projects are queued and none has been approved yet. Everything unreviewed is discovery metadata imported from GitHub — listed, not vetted.
Ranked by total stars at the last daily refresh. This is not stars earned during the selected period. Star counts in the AI category are known to be inflated by fake-star campaigns, so activity and license are shown first.
6 projects matching “on-device”
turbo-fieldfare
drumih/turbo-fieldfareGemma 4 26B-A4B inference in ~2 GB of RAM on any M-series MacBook
Pushed 3 days ago.Apache License 2.0SwiftDeveloper tools7.3k+94
openmed
maziyarpanahi/openmedLocal-first healthcare AI: clinical NER and HIPAA PII de-identification on hardware you control. 2,200+ medical models, 35 model-backed PII languages, and Python, MLX, Android and browser runtimes. Apache-2.0 SDK.
Pushed yesterday.Apache License 2.0PythonDeveloper tools5.8k
FedML
FedML-AI/FedMLFEDML - The unified and scalable ML library for large-scale distributed training, model serving, and federated learning. FEDML Launch, a cross-cloud scheduler, further enables running any AI jobs on any GPU cloud or on-premise cluster. Built on this library, TensorOpera AI (https://TensorOpera.ai) is your generative AI platform at scale.
Pushed today.Apache License 2.0PythonDeveloper tools4.1k
claude-code-local
nicedreamzapp/claude-code-localRun Claude Code 100% on-device with local AI on Apple Silicon. MLX-native Anthropic-API server. 6 fighters incl. Muse-Glimmer 30B (now multimodal — reads images, abliterated), Gemma 4 31B, Qwen 3.5 122B (65 tok/s), DeepSeek V4 Flash (1M ctx). Private, offline, airgap-ready. Built for NDA / legal / healthcare workflows.
Pushed today.MIT LicensePythonDeveloper tools3.5k+24
laya-coreml
mizorewww/laya-coremlLocal Laya typed decisions on Apple Core ML and Neural Engine. Validated ports, ~5 ms short decisions on M3 Max, reproducible speed and energy benchmarks.
Pushed 15 days ago.Apache License 2.0PythonDeveloper tools1.6k