Zana za CLI za kuandika code

Coding tool inahitaji mipangilio mitatu ili ifanye kazi na Shannon: base URL ya API, API key yako, na model id ya Shannon. Ukurasa huu unatoa mipangilio kamili ya Claude Code na Codex CLI, unaeleza jinsi model ya coding shannon-coder-1 inavyohesabiwa, na unaorodhesha makosa ambayo tools hizi hukutana nayo.

Kila tool inatumia endpoint ipi

Endpoint zote tatu zinakubali API key ile ile, kama Authorization: Bearer au kama x-api-key. Headers anthropic-version, anthropic-beta na openai-beta ambazo tools hizi hutuma zinakubaliwa.

Chagua model ya Shannon

Kila ombi lazima litaje model id ambayo Shannon inachapisha. Coding tool hutuma jina la model yake chaguo-msingi isipoambiwa vinginevyo, na jina ambalo Shannon haichapishi hujibiwa kwa 400 na ujumbe unknown model: <name>. Hakuna jina lingine linalolinganishwa na model ya Shannon, kwa hiyo weka model kwenye tool kabla ya ombi la kwanza.

GET /v1/models inaorodhesha kila id pamoja na dirisha lake la muktadha. Haihitaji key. Model na bei

Coding tools huongoza model kwa system prompt ndefu. System prompt inatumiwa na shannon-coder-1, shannon-1.6-lite, shannon-1.6-pro, shannon-2-lite, shannon-2-pro na model za open-weight zinazopangishwa, kwa hiyo chagua mojawapo kwa coding tool.

Claude Code

Claude Code husoma muunganisho wake kutoka environment variables. Ziweke kwenye shell inayoianzisha, au kwenye block ya env ya ~/.claude/settings.json ili ziendelee kwa kila session.

# 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"
  }
}

• Weka variables zote tatu za ANTHROPIC_DEFAULT_…_MODEL. Subagents na baadhi ya amri huomba model kwa jina lake la ndani, na kila jina lazima litatuliwe kuwa id ya Shannon.
• Mfano unaelekeza maombi ya usuli kwa shannon-1.6-lite, ambayo huhesabiwa kwa tokens. Vichwa vya session na maombi kama hayo basi havitumii wito wa Coder.
• Claude Code huuliza POST /v1/messages/count_tokens ukubwa wa muktadha. Endpoint hiyo huhesabu kwa model za open-weight zinazopangishwa. Kwa id ya Shannon hujibu 400; Claude Code huchukulia kuhesabu tokens kama hiari na hukadiria ukubwa yenyewe. Kuhesabu tokens

Ili kukagua anwani, key na model id kabla ya kuanzisha tool, tuma ombi moja kwa mkono:

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 huongea na Responses API. Eleza Shannon kama model provider kwenye ~/.codex/config.toml, chagua model id ya Shannon hapo, na uweke API key yako kwenye environment variable ambayo provider inataja.

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 haihifadhi mazungumzo kwenye seva: kila ombi hubeba mazungumzo kwenye input. previous_response_id inakubaliwa kwa ulinganifu.
• Function tools (type: "function") hutolewa kwa model kwenye /v1/responses.
• Kwenye /v1/responses, shannon-coder-1 huhesabiwa kwa tokens kutoka kwenye salio lako, si kwa wito wa Coder.

Ukaguzi ule ule kwa endpoint ya Responses:

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."
  }'

Model ya coding na jinsi inavyohesabiwa

shannon-coder-1 ni model ya Shannon kwa coding agents. GET /v1/models inairipoti ikiwa na dirisha la muktadha la tokens 131,072, wito wa tools, reasoning na input ya maandishi. Inahesabiwa kwa njia mojawapo kati ya mbili, kulingana na endpoint ambayo tool inatumia.

Wito ni ombi moja, bila kujali ukubwa wake. Huhesabiwa ombi linapokubaliwa, kabla model haijajibu. Session ya agent hutuma ombi moja kwa kila hatua inayochukua, kwa hiyo kazi ndefu hutumia wito mwingi. Kiasi hutegemea mpango wako na huanza upya kila saa 4, saa 00:00, 04:00, 08:00, 12:00, 16:00 na 20:00 UTC.

Ukurasa wa mipango unaonyesha namba za sasa. Mipango

Wito wa dirisha unapoisha, ombi hujibiwa kwa 429. Kwenye /v1/messages kosa lile lile huja kwa umbo la Anthropic.

{
  "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"
  }
}

Muktadha kuzidi

Mazungumzo yanapokuwa makubwa kuliko dirisha la muktadha la model, /v1/chat/completions na /v1/messages hujibu kwa 400, aina invalid_request_error, code context_length_exceeded na ujumbe unaoanza na prompt is too long. Coding tools husoma kosa hili ili kufupisha session na kutuma ombi tena. Linatolewa na shannon-coder-1, shannon-1.6-lite na shannon-1.6-pro.

{
  "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
  }
}

Ujumbe ni prompt is too long: <sent> tokens > <limit> maximum. Namba hizo mbili zisipojulikana ni prompt is too long. Please compact the conversation and try again.

Kwenye stream hali tayari ni 200, kwa hiyo kosa hufika kama tukio na stream kisha hufungwa kwa njia ya kawaida:

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"}

Kile agent loop inaweza kutegemea

Utatuzi wa matatizo

Key yako na salio lako viko kwenye ukurasa wa Keys na matumizi. Keys na matumizi