Deployment frameworks, AI coding harnesses, and ecosystem resources for working with LongCat models. Each entry includes verified source links and documented integration details.
Independent third-party resource. Not affiliated with, endorsed by, or sponsored by Meituan or LongCat. See the disclaimer and about page.
High-performance LLM serving framework with GPU and NPU support for LongCat models. SGLang PR #30042 adds dedicated LongCat-2.0 inference with tensor parallelism and expert parallelism. A separate NPU fork (SGLang-FluentLLM) provides optimized inference on domestic AI accelerators.
High-throughput LLM serving engine. LongCat models including LongCat-Flash-Chat-FP8 are directly loadable via vLLM with OpenAI-compatible API. HuggingFace model cards provide vLLM serve commands for quick deployment.
Anthropic's agentic coding CLI tool. LongCat-2.0 is documented as compatible through the Anthropic-compatible API endpoint at api.longcat.chat/anthropic. On OpenRouter, LongCat-2.0 ranked second in monthly call volume among all models used through Claude Code, behind only Claude Opus 4.8.
Open-source personal AI assistant with a dedicated LongCat provider plugin. Supports thinking mode mapping and streaming. Documentation recommends routing self-hosted LongCat models through vLLM or SGLang providers. LongCat-2.0 ranked third in monthly call volume through OpenClaw on OpenRouter.
Third-party API routing platform where LongCat-2.0 was initially listed anonymously as Owl Alpha. Provides unified API access to LongCat models alongside other LLM providers. LongCat-2.0 ranked among the top three models globally by daily token volume during its anonymous period.
LongCat models can be loaded directly via Hugging Face Transformers with trust_remote_code=True. Supports AutoModelForCausalLM for language models. BF16, FP8, and INT8 precision variants available.