| name | libertai-api |
| description | Use LibertAI's OpenAI-compatible confidential inference API for private, decentralized AI model calls (chat completions, embeddings, image generation) with no data logging and verifiable privacy via Trusted Execution Environments. |
LibertAI API Skill
This skill teaches agents how to call LibertAI's confidential inference API. LibertAI is an OpenAI-compatible, pay-per-token API running open-source models on decentralized infrastructure (Aleph Cloud) with TEE-backed privacy and no data logging.
When to use this skill
- You need a private, decentralized LLM backend instead of OpenAI/Anthropic.
- You want OpenAI-compatible calls with no vendor lock-in.
- Your workload must not be logged, trained on, or exposed to a centralized provider.
- You need open-source models (e.g. GLM-5.2, Qwen3 Coder) with transparent pricing.
Base URL and authentication
- Base URL:
https://api.libertai.io/v1
- Auth:
Authorization: Bearer <LIBERTAI_API_KEY> (create keys at https://console.libertai.io)
- Alternative auth: You can also use x402 payments, using the
@libertai/x402 npm package in TypeScript or libertai-x402 pypi package in Python
- Compatibility: Drop-in replacement for the OpenAI SDK. Set
baseURL / base_url and api_key.
Chat completions (OpenAI-compatible)
curl https://api.libertai.io/v1/chat/completions \
-H "Authorization: Bearer $LIBERTAI_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "glm-5.2",
"messages": [
{"role": "user", "content": "Say hello"}
]
}'
Python (OpenAI SDK):
from openai import OpenAI
client = OpenAI(
base_url="https://api.libertai.io/v1",
api_key=os.environ["LIBERTAI_API_KEY"],
)
resp = client.chat.completions.create(
model="glm-5.2",
messages=[{"role": "user", "content": "Say hello"}],
)
print(resp.choices[0].message.content)
Models
Current text models include glm-5.2 (flagship), qwen3.5-122b, gemma-4-31b, deepseek-v4-flash, qwen3-coder, and hermes-3-8b-tee (TEE-served). Image generation and search are also available. See the live list and pricing at https://docs.libertai.io/apis/text/#pricing.
References
