Kayan CLI na rubuta code

Kayan rubuta code suna buƙatar saituna uku don aiki da Shannon: base URL na API, maɓallin API ɗinku, da model id na Shannon. Wannan shafin yana ba da ainihin saitunan Claude Code da Codex CLI, yana bayanin yadda ake ƙirga model na rubuta code shannon-coder-1, kuma yana lissafa kurakuran da waɗannan tools ke haɗuwa da su.

Wane endpoint kowane tool ke amfani da shi

Dukkan endpoints uku suna karɓar maɓallin API iri ɗaya, a matsayin Authorization: Bearer ko x-api-key. Ana karɓar headers anthropic-version, anthropic-beta da openai-beta da waɗannan tools ke aikawa.

Zaɓi model na Shannon

Dole kowace request ta ambaci model id da Shannon ke wallafawa. Tool na rubuta code yana aika sunan model ɗinsa na asali sai dai idan kun gaya masa daban, kuma sunan da Shannon ba ya wallafawa ana amsa shi da 400 da saƙon unknown model: <name>. Babu wani suna da aka haɗa da model na Shannon, don haka ku saita model a cikin tool kafin request na farko.

GET /v1/models yana lissafa kowane id tare da context window ɗinsa. Ba ya buƙatar maɓalli. Models & farashi

Kayan rubuta code suna jagorantar model da dogon system prompt. shannon-coder-1, shannon-1.6-lite, shannon-1.6-pro, shannon-2-lite, shannon-2-pro da hosted open-weight models ne ke aiwatar da system prompt, don haka ku zaɓi ɗaya daga cikinsu don tool na rubuta code.

Claude Code

Claude Code yana karanta haɗinsa daga environment variables. Ku saita su a shell da ke ƙaddamar da shi, ko a block na env na ~/.claude/settings.json don riƙe su a kowane 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"
  }
}

• Ku saita dukkan variables uku na ANTHROPIC_DEFAULT_…_MODEL. Subagents da wasu umarni suna neman model da sunansa na ciki, kuma kowane suna dole ya zama id na Shannon.
• Samfurin yana nuna requests na baya-baya zuwa shannon-1.6-lite, wanda ake ƙirga a tokens. Sunayen session da irin waɗannan requests sai ba sa amfani da kiran Coder.
• Claude Code yana tambayar POST /v1/messages/count_tokens girman context. Endpoint ɗin yana ƙirga ga hosted open-weight models. Da id na Shannon yana amsawa da 400; Claude Code yana ɗaukar ƙirgan tokens a matsayin zaɓi kuma yana kimanta girman da kansa. Ƙidayar tokens

Don duba adireshin, maɓalli da model id kafin ku fara tool, ku aika request ɗaya da hannu:

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 yana magana da Responses API. Ku bayyana Shannon a matsayin model provider a ~/.codex/config.toml, ku zaɓi model id na Shannon a can, kuma ku sa maɓallin API ɗinku a environment variable da provider ya ambata.

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 ba ya riƙe tattaunawa a server: kowace request tana ɗauke da tattaunawar a input. Ana karɓar previous_response_id don jituwa.
• Ana ba model function tools (type: "function") a /v1/responses.
• A /v1/responses, ana ƙirga shannon-coder-1 a tokens daga balance ɗinku, ba a kiran Coder ba.

Duba iri ɗaya ga endpoint na 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 na rubuta code da yadda ake ƙirga shi

shannon-coder-1 shi ne model na Shannon don coding agents. GET /v1/models yana bayyana shi da context window na tokens 131,072, kiran tool, reasoning da rubutu a input. Ana ƙirga shi ta hanya ɗaya daga biyu, bisa endpoint da tool ke amfani da shi.

Kira request ɗaya ne, komai girmansa. Ana ƙirga shi lokacin da aka karɓi request, kafin model ya amsa. Session na agent yana aika request ɗaya don kowane mataki da ya ɗauka, don haka dogon aiki yana amfani da kira da yawa. Allowance ya dogara da plan ɗinku kuma yana farawa kowane awa 4, a 00:00, 04:00, 08:00, 12:00, 16:00 da 20:00 UTC.

Shafin plan yana nuna lambobin yanzu. Plans

Idan an gama kiran taga, ana amsa request da 429. A /v1/messages kuskuren yana zuwa a siffar 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"
  }
}

Cikar context

Idan tattaunawa ta fi context window na model girma, /v1/chat/completions da /v1/messages suna amsawa da 400, nau'in invalid_request_error, code context_length_exceeded da saƙon da ya fara da prompt is too long. Kayan rubuta code suna karanta wannan kuskuren don matse session su sake aika request. shannon-coder-1, shannon-1.6-lite da shannon-1.6-pro ne ke samar da shi.

{
  "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
  }
}

Saƙon shi ne prompt is too long: <sent> tokens > <limit> maximum. Idan ba a san lambobin biyu ba, shi ne prompt is too long. Please compact the conversation and try again.

A stream matsayin ya riga ya zama 200, don haka kuskuren yana zuwa a matsayin event sannan stream ya rufe ta hanyar da ta saba:

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"}

Abin da agent loop zai iya dogara da shi

Magance matsaloli

Maɓallinku da balance ɗinku suna a shafin Maɓallai & amfani. Maɓallai & amfani