README: focus on the model, remove API endpoint

README.md

@@ -26,7 +26,6 @@ security workflows and autonomous security agents.
 - **Unrestricted** for legitimate security use cases.
 - **27B parameters**, long context (up to 256k tokens), native tool-calling.
 - Optimized for pentest agents and security copilots.
-- Deployed and maintained by **Nesilabs** — infrastructure 100% in the EU (GDPR).
 
 ## Model details
 | | |
@@ -39,19 +38,20 @@ security workflows and autonomous security agents.
 | Serving | SGLang / vLLM compatible |
 
 ## Usage
-Nesi1 is available as a managed service via the **Nesilabs API**
-(`api.nesilabs.com`). Contact Nesilabs for access credentials.
+Nesi1 runs on any OpenAI-compatible inference server (SGLang, vLLM). Example
+with SGLang:
 
-```python
-from openai import OpenAI
-client = OpenAI(base_url="https://api.nesilabs.com/v1", api_key="YOUR_KEY")
-resp = client.chat.completions.create(
-    model="Nesi1",
-    messages=[{"role": "user", "content": "Explain how to test for SQL injection."}],
-)
-print(resp.choices[0].message.content)
+```bash
+python3 -m sglang.launch_server \
+  --model-path nesilabs/Nesi1 \
+  --served-model-name Nesi1 \
+  --context-length 262144 \
+  --tool-call-parser qwen3_coder \
+  --trust-remote-code
 ```
 
+Then query it like any OpenAI chat endpoint (`/v1/chat/completions`).
+
 ## ⚠️ Legal Notice / Disclaimer
 
 **Nesi1 is intended EXCLUSIVELY for security professionals, for lawful and