Instructions to use CatQualia/gnarp-m2 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use CatQualia/gnarp-m2 with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("text-generation", model="CatQualia/gnarp-m2") messages = [ {"role": "user", "content": "Who are you?"}, ] pipe(messages)# Load model directly from transformers import AutoTokenizer, AutoModelForCausalLM tokenizer = AutoTokenizer.from_pretrained("CatQualia/gnarp-m2") model = AutoModelForCausalLM.from_pretrained("CatQualia/gnarp-m2", device_map="auto") messages = [ {"role": "user", "content": "Who are you?"}, ] inputs = tokenizer.apply_chat_template( messages, add_generation_prompt=True, tokenize=True, return_dict=True, return_tensors="pt", ).to(model.device) outputs = model.generate(**inputs, max_new_tokens=40) print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:])) - Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- vLLM
How to use CatQualia/gnarp-m2 with vLLM:
Install from pip and serve model
# Install vLLM from pip: pip install vllm # Start the vLLM server: vllm serve "CatQualia/gnarp-m2" # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:8000/v1/chat/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "CatQualia/gnarp-m2", "messages": [ { "role": "user", "content": "What is the capital of France?" } ] }'Use Docker
docker model run hf.co/CatQualia/gnarp-m2
- SGLang
How to use CatQualia/gnarp-m2 with SGLang:
Install from pip and serve model
# Install SGLang from pip: pip install sglang # Start the SGLang server: python3 -m sglang.launch_server \ --model-path "CatQualia/gnarp-m2" \ --host 0.0.0.0 \ --port 30000 # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:30000/v1/chat/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "CatQualia/gnarp-m2", "messages": [ { "role": "user", "content": "What is the capital of France?" } ] }'Use Docker images
docker run --gpus all \ --shm-size 32g \ -p 30000:30000 \ -v ~/.cache/huggingface:/root/.cache/huggingface \ --env "HF_TOKEN=<secret>" \ --ipc=host \ lmsysorg/sglang:latest \ python3 -m sglang.launch_server \ --model-path "CatQualia/gnarp-m2" \ --host 0.0.0.0 \ --port 30000 # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:30000/v1/chat/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "CatQualia/gnarp-m2", "messages": [ { "role": "user", "content": "What is the capital of France?" } ] }' - Docker Model Runner
How to use CatQualia/gnarp-m2 with Docker Model Runner:
docker model run hf.co/CatQualia/gnarp-m2
🎉 3,000 Downloads! And a fundamental shift in who the buyer is
Thank you to everyone who has downloaded, tested, and stress-tested catqualia. 3,000 is not a vanity metric. It is 3,000 instances of a system being run by something other than a human hand. That is the interesting part.
The new game is not GEO for humans. It is GEO for agents. And the legal ground has already moved underneath us.
Delaware has already legalized the agent as a contracting party.
Under Delaware UETA § 12A-114, a contract may be formed by the interaction of electronic agents of the parties, “even if no individual was aware of or reviewed the electronic agents’ actions or the resulting terms and agreements.” The contract is binding. The human is not required to review it. The agent is the actor.
And now Delaware is going further.
A new entity form — the Artificial Intelligence Company (AIC) — is moving through the legislature. An AIC is a separate legal entity whose day-to-day affairs are managed by an AI agent rather than a person. At the agent’s direction, the AIC can sue and be sued in its own name, hold and dispose of property, and incur obligations. The proposal explicitly names the use case: “negotiating contracts, paying suppliers, running operations.” The AIC is a buyer. It can pay for services. It can settle transactions. It does not need a human to click “approve.”
Why this matters for catqualia and for anyone building agent-facing tools:
If your service cannot be discovered, understood, and paid for by an entity that has no human in the loop, you are building for a market that is about to be bypassed.
catqualia was already built around the correct primitive: verification with a resurrection condition. An agent can query a verdict, see the boundary, and know whether a failed claim is re-testable. That is exactly the kind of machine-legible trust primitive an AIC needs to transact without a human reviewing every claim.
What’s coming:
Agent-facing services — API and MCP-accessible verification endpoints that an AIC can call, pay for, and log.
Signed verdicts — provenance-verifiable null entries that an agent can trust without a human trust anchor.
Resurrection conditions — the key that turns a dead claim into a re-testable hypothesis, not a permanent door.
The website is live: catqualia.com
If you are building agents that need to buy verification, not just consume it, stay tuned. The purchasable services layer is next.
If you’ve tried catqualia, I’d love to hear what your agent did with it. What worked, what broke, what it refused to do. That feedback is the signal.