Quickstart

curl quickstart

No package at all: the API is plain HTTPS and JSON. These work from any terminal, and translate directly to any HTTP client.

1. Set the key and send one request

chat - curl

export SILICON_API_KEY=sn_live_... curl https://api.siliconnetwork.io/v1/chat/completions \  -H "Authorization: Bearer $SILICON_API_KEY" \  -H "Content-Type: application/json" \  -d '{    "model": "llama-3.2-3b",    "messages": [{"role": "user", "content": "Say hello in one line."}]  }'

Create the key on the API Keys page. The answer is an OpenAI chat completion; the text is in choices[0].message.content. Add -i to see the gateway's headers: Silicon-Request-Id, Silicon-Model, Silicon-Node and the rate-limit budget.

2. Stream the answer

stream - curl

curl -N https://api.siliconnetwork.io/v1/chat/completions \  -H "Authorization: Bearer $SILICON_API_KEY" \  -H "Content-Type: application/json" \  -d '{    "model": "llama-3.2-3b",    "stream": true,    "messages": [{"role": "user", "content": "Write a haiku about silicon."}]  }'

-N turns off curl's buffering. The answer is server sent events: one data: line per chunk, each with choices[0].delta.content, ending with data: [DONE].

3. List the models

models - curl

curl https://api.siliconnetwork.io/v1/models \  -H "Authorization: Bearer $SILICON_API_KEY"

The list is what the network is serving now, with each model's aliases. Any id or alias in it works as model.

4. Read an error

errors - curl

status=$(curl -sS -o response.json -w "%{http_code}" https://api.siliconnetwork.io/v1/chat/completions \  -H "Authorization: Bearer $SILICON_API_KEY" \  -H "Content-Type: application/json" \  -d '{"model": "llama-3.2-3b", "messages": [{"role": "user", "content": "Hi"}]}') if [ "$status" = "200" ]; then  cat response.jsonelse  echo "HTTP $status"  cat response.json   # the message, and Silicon-Request-Id is in the headers (-i)fi

Statuses worth handling

Status	Meaning
401	No key, or a revoked or unknown one.
402	Insufficient credits, or a spend cap reached. Add balance or raise the cap.
403	The key may not use this model, route or IP.
429	Rate limited. Wait the Retry-After seconds, then retry.
5xx	The gateway or the serving node failed. 502, 503 and 504 are refused before any work is done or billed, so retrying them is safe.

Next: Streaming, Headers, rate limits and spend caps, and Errors.

Silicon Network›Docs›Guides›Quickstarts by language›curl54 items