Terminal & curl

Discover models with GET /v1/models, call OpenAI-compatible endpoints with curl, stream responses, and use the Python SDK against api.undersky.ai.

Use standard HTTP tools against https://api.undersky.ai. Create API keys, top up credits, and manage your account in the Dashboard.

UnderSky CLI (account tools)

Install the undersky-cli / usai binaries for login, doctor checks, and control-plane helpers:

curl -fsSL https://raw.githubusercontent.com/leaperone/UnderSky/main/scripts/install-cli.sh | sh

Update with undersky-cli upgrade, or re-run the installer. Prebuilt archives ship on GitHub Releases tagged cli-v* (linux/macOS, x64 and arm64) with SHA256SUMS. If the GitHub repository is private, set GH_TOKEN (or GITHUB_TOKEN) with contents:read for install and upgrade.

Set your API key

Store your model API key in the shell for repeatable examples:

export UNDERSKY_API_KEY="sk-your-api-key"

Never commit real API keys. Use Dashboard → Settings → API Keys to rotate compromised keys.

Discover models and endpoints

Public catalog endpoints do not require authentication:

curl -sS https://api.undersky.ai/v1/models | jq '.data[] | {id, owned_by}'
curl -sS https://api.undersky.ai/openapi.json | jq '.paths | keys[:10]'

Browse the full reference on undersky.ai/docs.

Chat completion

curl -sS https://api.undersky.ai/v1/chat/completions \
  -H "Authorization: Bearer $UNDERSKY_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-5-nano",
    "messages": [{"role": "user", "content": "Hello!"}]
  }' | jq '.choices[0].message.content'

Streaming

curl -N https://api.undersky.ai/v1/chat/completions \
  -H "Authorization: Bearer $UNDERSKY_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-5-nano",
    "stream": true,
    "messages": [{"role": "user", "content": "Count to three."}]
  }'

OpenAI SDK (Python)

pip install openai

from openai import OpenAI

client = OpenAI(
    api_key="sk-your-api-key",
    base_url="https://api.undersky.ai/v1",
)

response = client.chat.completions.create(
    model="gpt-5-nano",
    messages=[{"role": "user", "content": "Hello!"}],
)
print(response.choices[0].message.content)

Credits and usage

Account balance and usage history are available in Dashboard → Settings → Credit & Usage. Server integrations that already hold a scoped control token (lpc_…) can query balance with:

curl -sS https://api.undersky.ai/api/v1/credits/balance \
  -H "Authorization: Bearer lpc-your-control-token" | jq

Scripting tips

Pipe JSON through jq, capture HTTP status with -w "%{http_code}", and retry on 429 with backoff. Parse response bodies only; treat stderr from your shell separately from API error payloads.

On this page