CLI kodlash vositalari

Kodlash vositasi Shannon bilan ishlashi uchun uchta sozlama kerak: API'ning base URL'i, API kalitingiz va Shannon model id'si. Bu sahifa Claude Code va Codex CLI uchun aniq sozlamalarni beradi, shannon-coder-1 kodlash modeli qanday hisoblanishini tushuntiradi va bu vositalar duch keladigan xatolarni sanab beradi.

Har bir vosita qaysi endpointdan foydalanadi

Uchala endpoint ham bir xil API kalitni Authorization: Bearer yoki x-api-key sifatida qabul qiladi. Bu vositalar yuboradigan anthropic-version, anthropic-beta va openai-beta sarlavhalari qabul qilinadi.

Shannon modelini tanlash

Har bir so'rov Shannon e'lon qilgan model id'sini ko'rsatishi shart. Kodlash vositasi, aks holda aytmasangiz, o'zining standart model nomini yuboradi, Shannon e'lon qilmagan nomga esa 400 va unknown model: <name> xabari bilan javob beriladi. Boshqa hech bir nom Shannon modeliga moslanmaydi, shuning uchun birinchi so'rovdan oldin modelni vositada o'rnating.

GET /v1/models har bir id'ni kontekst oynasi bilan sanab chiqadi. Kalit kerak emas. Modellar va narxlar

Kodlash vositalari modelni uzun system prompt bilan boshqaradi. System promptni shannon-coder-1, shannon-1.6-lite, shannon-1.6-pro, shannon-2-lite, shannon-2-pro va xostingdagi open-weight modellar qo'llaydi, shuning uchun kodlash vositasi uchun shulardan birini tanlang.

Claude Code

Claude Code ulanishini muhit o'zgaruvchilaridan o'qiydi. Ularni uni ishga tushiradigan shellda yoki har bir sessiya uchun saqlash uchun ~/.claude/settings.json ning env blokida o'rnating.

# Install Claude Code
npm install -g @anthropic-ai/claude-code

# Where to send requests, and your Shannon API key
export ANTHROPIC_BASE_URL=https://api.shannon-ai.com
export ANTHROPIC_AUTH_TOKEN=YOUR_API_KEY

# Shannon model ids: the session model, and the model behind each built-in name
export ANTHROPIC_MODEL=shannon-coder-1
export ANTHROPIC_DEFAULT_OPUS_MODEL=shannon-coder-1
export ANTHROPIC_DEFAULT_SONNET_MODEL=shannon-coder-1
export ANTHROPIC_DEFAULT_HAIKU_MODEL=shannon-1.6-lite

# The context window of the session model
export CLAUDE_CODE_MAX_CONTEXT_TOKENS=131072

claude

{
  "env": {
    "ANTHROPIC_BASE_URL": "https://api.shannon-ai.com",
    "ANTHROPIC_AUTH_TOKEN": "YOUR_API_KEY",
    "ANTHROPIC_MODEL": "shannon-coder-1",
    "ANTHROPIC_DEFAULT_OPUS_MODEL": "shannon-coder-1",
    "ANTHROPIC_DEFAULT_SONNET_MODEL": "shannon-coder-1",
    "ANTHROPIC_DEFAULT_HAIKU_MODEL": "shannon-1.6-lite",
    "CLAUDE_CODE_MAX_CONTEXT_TOKENS": "131072"
  }
}

• Uchala ANTHROPIC_DEFAULT_…_MODEL o'zgaruvchisini ham o'rnating. Subagentlar va ba'zi buyruqlar modelni o'rnatilgan nomi bilan so'raydi va har bir nom Shannon id'siga aylanishi kerak.
• Namuna fon so'rovlarini token bilan hisoblanadigan shannon-1.6-litega yo'naltiradi. Shunda sessiya sarlavhalari va shunga o'xshash so'rovlar Coder chaqiruvlarini sarflamaydi.
• Claude Code kontekst o'lchamini POST /v1/messages/count_tokens dan so'raydi. Endpoint xostingdagi open-weight modellar uchun sanaydi. Shannon id'si bilan u 400 bilan javob beradi; Claude Code token sanashni ixtiyoriy deb hisoblaydi va o'lchamni o'zi baholaydi. Tokenlarni sanash

Vositani ishga tushirishdan oldin manzil, kalit va model id'sini tekshirish uchun bitta so'rovni qo'lda yuboring:

curl https://api.shannon-ai.com/v1/messages \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "shannon-1.6-lite",
    "max_tokens": 16,
    "messages": [{"role": "user", "content": "Say OK."}]
  }'

Codex CLI

Codex CLI Responses API bilan gaplashadi. Shannon'ni ~/.codex/config.toml da model provayderi sifatida tasvirlang, u yerda Shannon model id'sini tanlang va API kalitingizni provayder nomlagan muhit o'zgaruvchisiga qo'ying.

model = "shannon-coder-1"
model_provider = "shannon"
model_context_window = 131072

[model_providers.shannon]
name = "Shannon"
base_url = "https://api.shannon-ai.com/v1"
env_key = "SHANNON_API_KEY"
wire_api = "responses"

# Install Codex CLI
npm install -g @openai/codex

# Your Shannon API key, under the name given as env_key
export SHANNON_API_KEY=YOUR_API_KEY

# Start it; --model selects another Shannon id for one run
codex
codex --model shannon-1.6-pro

• /v1/responses serverda suhbatni saqlamaydi: har bir so'rov suhbatni inputda olib keladi. previous_response_id moslik uchun qabul qilinadi.
• Funksiya tool'lari (type: "function") /v1/responses da modelga taklif qilinadi.
• /v1/responses da shannon-coder-1 Coder chaqiruvlarida emas, balansingizdan token bilan hisoblanadi.

Responses endpointi uchun xuddi shu tekshiruv:

curl https://api.shannon-ai.com/v1/responses \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "shannon-1.6-lite",
    "input": "Say OK."
  }'

Kodlash modeli va u qanday hisoblanadi

shannon-coder-1 — kodlash agentlari uchun Shannon modeli. GET /v1/models uni 131,072 tokenli kontekst oynasi, tool chaqiruvlari, mulohaza va matn kirishi bilan ko'rsatadi. U vosita ishlatadigan endpointga qarab ikki usuldan biri bilan hisoblanadi.

Chaqiruv — bitta so'rov, o'lchamidan qat'i nazar. U so'rov qabul qilinganda, model javob berishidan oldin hisoblanadi. Agent sessiyasi qilgan har bir qadam uchun bitta so'rov yuboradi, shuning uchun uzoq vazifa ko'p chaqiruv sarflaydi. Limit rejangizga bog'liq va har 4 soatda, 00:00, 04:00, 08:00, 12:00, 16:00 va 20:00 UTC da qaytadan boshlanadi.

Joriy sonlar reja sahifasida ko'rsatilgan. Rejalar

Oyna chaqiruvlari tugagach, so'rovga 429 bilan javob beriladi. /v1/messages da xuddi shu xato Anthropic shaklida keladi.

{
  "error": {
    "type": "rate_limit_error",
    "message": "Shannon Coder call quota reached. Upgrade your plan at shannon-ai.com/plan"
  }
}

{
  "type": "error",
  "error": {
    "type": "rate_limit_error",
    "message": "Shannon Coder call quota reached. Upgrade your plan at shannon-ai.com/plan"
  }
}

Kontekstning to'lib ketishi

Suhbat modelning kontekst oynasidan katta bo'lganda /v1/chat/completions va /v1/messages 400, invalid_request_error turi, context_length_exceeded kodi va prompt is too long bilan boshlanadigan xabar bilan javob beradi. Kodlash vositalari bu xatoni sessiyani siqish va so'rovni qayta yuborish uchun o'qiydi. Uni shannon-coder-1, shannon-1.6-lite va shannon-1.6-pro hosil qiladi.

{
  "error": {
    "type": "invalid_request_error",
    "code": "context_length_exceeded",
    "message": "prompt is too long: 140211 tokens > 131072 maximum",
    "param": null
  }
}

{
  "type": "error",
  "error": {
    "type": "invalid_request_error",
    "code": "context_length_exceeded",
    "message": "prompt is too long: 140211 tokens > 131072 maximum",
    "param": null
  }
}

Xabar: prompt is too long: <sent> tokens > <limit> maximum. Ikkala son noma'lum bo'lganda: prompt is too long. Please compact the conversation and try again.

Oqimda holat allaqachon 200, shuning uchun xato hodisa sifatida keladi va oqim keyin odatdagidek yopiladi:

data: {"error":{"message":"prompt is too long: 140211 tokens > 131072 maximum","type":"invalid_request_error","code":"context_length_exceeded","param":null}}

data: {"id":"chatcmpl-…","object":"chat.completion.chunk","created":…,"model":"shannon-coder-1","choices":[{"index":0,"delta":{},"finish_reason":"stop"}],"usage":{…}}

data: [DONE]

event: error
data: {"type":"error","error":{"type":"invalid_request_error","message":"prompt is too long: 140211 tokens > 131072 maximum"}}

event: message_delta
data: {"type":"message_delta","delta":{"stop_reason":"end_turn","stop_sequence":null},"usage":{…}}

event: message_stop
data: {"type":"message_stop"}

Agent sikli nimaga tayanishi mumkin

Nosozliklarni bartaraf etish

Kalitingiz va balansingiz Keys & usage sahifasida. Keys & usage