Arcee AI

Founded 2023

Deploy and fine-tune open-weight language models via API or self-hosting

You can use Arcee AI to access, customize, and deploy open-weight language models for specialized enterprise applications. It provides pre-trained models like Trinity and Genesis-Science-1 designed for complex reasoning, everyday operational tasks, and scientific research. You can query models directly through a hosted API or download the weights to run on your own private infrastructure and edge devices. Teams can inspect architectures, fine-tune models on domain data, and run long-running autonomous agents using open-source frameworks without locking into closed proprietary vendors.

What Arcee AI does

Download open model weights for local executionSend text generation prompts through the hosted APIFine-tune pre-trained weights on custom datasetsDeploy small models to edge hardware and laptopsConfigure autonomous agent tasks using the NAC harnessEvaluate reasoning performance on complex logical queriesRun private offline inference inside secure enterprise clustersOpen-weight model family with inspectable and downloadable model weightsHosted API for sending inference requests without managing hardwareSupport for self-hosting models on private on-premise servers and edge hardwareTrinity Nano model designed for fast on-device inferenceTrinity Large Thinking model designed for deep multistep logical reasoningNAC open-source harness for coordinating long-running autonomous agentsPre-trained Genesis-Science-1 model tuned specifically for scientific researchDomain-specific fine-tuning options on private company datasets