Chat Completions

POST https://api.shannon-ai.com/v1/chat/completions

Request kacha nta bụ model id na otu user message.

from openai import OpenAI

client = OpenAI(
    api_key="YOUR_API_KEY",
    base_url="https://api.shannon-ai.com/v1",
)

response = client.chat.completions.create(
    model="shannon-3",
    messages=[{"role": "user", "content": "Say hello in one sentence."}],
)

print(response.choices[0].message.content)

import OpenAI from "openai";

const client = new OpenAI({
  apiKey: "YOUR_API_KEY",
  baseURL: "https://api.shannon-ai.com/v1",
});

const response = await client.chat.completions.create({
  model: "shannon-3",
  messages: [{ role: "user", content: "Say hello in one sentence." }],
});

console.log(response.choices[0].message.content);

curl https://api.shannon-ai.com/v1/chat/completions \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "shannon-3",
    "messages": [{"role": "user", "content": "Say hello in one sentence."}]
  }'

Azịza bụ otu JSON object:

{
  "id": "chatcmpl-5f0c1e7a9b3d4c62a8e1f07d2b46c9a3",
  "object": "chat.completion",
  "created": 1791625200,
  "model": "shannon-3",
  "choices": [
    {
      "index": 0,
      "message": {
        "role": "assistant",
        "content": "Hello, it is good to meet you.",
        "reasoning_content": "The user wants a greeting in one sentence. Keep it short and friendly."
      },
      "finish_reason": "stop"
    }
  ],
  "usage": {
    "prompt_tokens": 1184,
    "completion_tokens": 46,
    "total_tokens": 1230
  }
}

Headers

Request headers

Reply headers

Request fields

Naanị messages ka a chọrọ. Ogidi Ndị na-etinye ya na-akpọ aha models ndị field na-agbanwe azịza na ha. Hosted open-weight models bụ ids iri na abụọ nke ndepụta model; ezinụlọ Shannon 3 bụ shannon-3, shannon-3-pro, shannon-3.1 na shannon-3.1-pro. Models & ọnụahịa

A na-anabata fields OpenAI ndị ọzọ, dị ka n, user, stream_options, parallel_tool_calls, presence_penalty, frequency_penalty, logit_bias, logprobs, metadata, store na prompt_cache_key, ka koodu client dị adị na-agba n'enweghị mgbanwe. Ha anaghị agbanwe azịza: mgbe niile enwere otu choice, ma stream na-agwụ mgbe niile na usage.

Field nwere JSON type ezighi ezi, dịka "max_tokens": "100", na-eweghachi 422. Request na-enweghị messages na-eme otu ihe ahụ.

Tools, structured output, reasoning na web search nke ọ bụla nwere peeji nke ya: Oku ọrụ, Mmepụta ahaziri, Reasoning effort, Chọọ weebụ.

Request nwere nhọrọ

Request a na-esetịpụ system message, sampling fields na reasoning effort. Ọ na-eji hosted open-weight model, nke na-etinye ha niile.

from openai import OpenAI

client = OpenAI(
    api_key="YOUR_API_KEY",
    base_url="https://api.shannon-ai.com/v1",
)

response = client.chat.completions.create(
    model="DeepSeek-V4-Flash-0731-W4A16-AUTOROUND-REAP",
    messages=[
        {"role": "system", "content": "You are a physics teacher. Answer in two sentences."},
        {"role": "user", "content": "Why is the sky blue?"},
    ],
    max_tokens=512,
    temperature=0.3,
    top_p=0.9,
    seed=7,
    stop=["\n\n"],
    reasoning_effort="low",
)

message = response.choices[0].message
print(message.reasoning_content)  # the reasoning
print(message.content)            # the answer
print(response.usage)

import OpenAI from "openai";

const client = new OpenAI({
  apiKey: "YOUR_API_KEY",
  baseURL: "https://api.shannon-ai.com/v1",
});

const response = await client.chat.completions.create({
  model: "DeepSeek-V4-Flash-0731-W4A16-AUTOROUND-REAP",
  messages: [
    { role: "system", content: "You are a physics teacher. Answer in two sentences." },
    { role: "user", content: "Why is the sky blue?" },
  ],
  max_tokens: 512,
  temperature: 0.3,
  top_p: 0.9,
  seed: 7,
  stop: ["\n\n"],
  reasoning_effort: "low",
});

const message = response.choices[0].message;
console.log(message.reasoning_content); // the reasoning
console.log(message.content);           // the answer
console.log(response.usage);

curl https://api.shannon-ai.com/v1/chat/completions \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "DeepSeek-V4-Flash-0731-W4A16-AUTOROUND-REAP",
    "messages": [
      {"role": "system", "content": "You are a physics teacher. Answer in two sentences."},
      {"role": "user", "content": "Why is the sky blue?"}
    ],
    "max_tokens": 512,
    "temperature": 0.3,
    "top_p": 0.9,
    "seed": 7,
    "stop": ["\n\n"],
    "reasoning_effort": "low"
  }'

Azịza nwere otu udi dị ka n'elu. usage ya na-agbakwunye nkọwa abụọ na hosted open-weight models: prompt tokens e sitere na cache gụọ na tokens e jiri maka reasoning.

{
  "usage": {
    "prompt_tokens": 31,
    "completion_tokens": 62,
    "total_tokens": 93,
    "prompt_tokens_details": {
      "cached_tokens": 0
    },
    "completion_tokens_details": {
      "reasoning_tokens": 21
    }
  }
}

Ogologo output

max_tokens na-eme ihe abụọ. Nke mbụ, ọ bụ ọnụọgụ tokens a na-ewepụta n'ime balance gị mgbe request na-amalite. Mgbe azịza zuru ezu, a na-eji tokens request ji dochie ọnụọgụ ahụ. Ọ bụrụ na max_tokens karịrị ihe fọdụrụ na balance gị, request na-eweghachi 429 Quota exceeded ọbụna mgbe azịza n'onwe ya ga-ekwere. Zipu max_tokens dị ala iji wepụta obere.

A na-agụ shannon-coder-1 n'ụzọ dị iche n'endpoint a: request ọ bụla bụ otu n'ime oku Shannon Coder nke plan gị, ma a naghị ewepụta tokens maka ya. Limits na balance

Nke abụọ, ọ na-eguzobe ogologo azịza na models ndị a:

Na-enweghị max_tokens ma ọ bụ max_completion_tokens, uru bụ 4,096. Na shannon-coder-1 ọ bụ 65,536.

Messages

Ozi ọ bụla bụ object nwere role na content. content bụ string, ma ọ bụ array nke akụkụ mgbe ozi na-ebu karịa text.

Na id nke ezinụlọ Shannon 3, tinye ntụziaka ndị ga-adịgide n'ime ozi user.

Na ọkwa Shannon request na-enweghị user text na enweghị tools na-eweghachi 400 No user message provided.

Content parts

Nha, oke na ndepụta zuru ezu nke udi nwere peeji nke ha. Images na files

Azịza object

Ihe mere azịza ji kwụsị

Azịza na-abụghị streamed na-akọ stop ma ọ bụ tool_calls.

Usage

Na hosted open-weight models, prompt_tokens bụ ozi gị na nkọwa tool agụrụ site na tokenizer nke model n'onwe ya, gbakwunyere tokens nke foto ọ bụla. Endpoints ịgụ token na-eweghachi otu nọmba ahụ tupu ị zipụ. Ịgụ tokens

Na ọkwa Shannon, prompt_tokens na-agụ ihe niile model gụrụ ide azịza, yabụ ọ karịrị text nke ozi gị naanị.

Streaming

Mgbe stream bụ true azịza na-abịa dị ka chat.completion.chunk events ma na-agwụ na data: [DONE]. Chunk ikpeazụ tupu ya na-ebu finish_reason na usage; achọghị stream_options. Udi chunk, ahịrị keep-alive na njehie n'ime stream nwere peeji nke ha. Nzipu ozugbo

Njehie

Njehie bụ JSON object nwere member error. A na-agba nlele n'usoro a: API key, request body, model id, mgbe ahụ balance. Tebụl a depụtara ihe endpoint a na-eweghachi n'ọtụtụ oge. Ndepụta zuru ezu, na ihe a ga-anwa ọzọ, nwere peeji nke ya. Njikwa njehie

{
  "error": {
    "type": "invalid_request_error",
    "message": "unknown model: no-such-model"
  }
}