llm-runcrate

LLM plugin for Runcrate: 200+ open models (DeepSeek, Qwen, Llama, Kimi, gpt-oss and more) behind one OpenAI-compatible API, billed per token from one prepaid balance.

Install

llm install llm-runcrate

Configure

Create an API key (docs), then:

llm keys set runcrate
# or: export RUNCRATE_API_KEY=...

Use

llm models list | grep runcrate/
llm -m runcrate/deepseek-ai/DeepSeek-V3.2 "Explain LoRA fine-tuning in one sentence"
llm -m runcrate/Qwen/Qwen3-VL-8B-Instruct "Describe this" -a photo.jpg
llm chat -m runcrate/moonshotai/Kimi-K2.6

Models that support tools and JSON schemas work with --tool/-T and --schema. The model list comes from https://api.runcrate.ai/v1/models and is cached for a day; refresh it with llm runcrate refresh.

Python:

import llm

model = llm.get_model("runcrate/deepseek-ai/DeepSeek-V3.2")
print(model.prompt("Name three uses for an H100").text())

Models and per-token prices: runcrate.ai/inference/pricing. The same account rents GPUs by the hour, and the remote MCP server https://mcp.runcrate.ai/mcp (OAuth sign-in, no key) lets agents launch them (setup).

Development

uv run --group test pytest

License

MIT, Aeonmind LLC.