Tlola ho ea ho litaba
Litšoantšo le lifaele

Litšoantšo le lifaele

Romela litšoantšo le molaetsa: hore na ke li-model life tse li amohelang, mebopeho e amohelwang le meeli ea eona.

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

Setšoantšo se tsamaea ka hare ho molaetsa oa user, e le karolo e le 'ngoe ea content ea oona. Mohlala o bala faele ea lehae mme o e romela e le data URL.

import base64
from openai import OpenAI

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

with open("photo.jpg", "rb") as f:
    image = base64.b64encode(f.read()).decode()

response = client.chat.completions.create(
    model="shannon-3",
    messages=[{
        "role": "user",
        "content": [
            {"type": "text", "text": "What is in this picture?"},
            {"type": "image_url", "image_url": {"url": "data:image/jpeg;base64," + image}},
        ],
    }],
)

print(response.choices[0].message.content)
200 JSON
{
  "id": "chatcmpl-8d1f3a5c7e9b4d2f6a8c0e2b4d6f8a1c",
  "object": "chat.completion",
  "created": 1791590400,
  "model": "shannon-3",
  "choices": [
    {
      "index": 0,
      "message": {
        "role": "assistant",
        "content": "A grey cat asleep on a windowsill, with a potted fern beside it.",
        "reasoning_content": null
      },
      "finish_reason": "stop"
    }
  ],
  "usage": {
    "prompt_tokens": 286,
    "completion_tokens": 19,
    "total_tokens": 305
  }
}

Li-model life tse amohelang litšoantšo

Li-model E bala
shannon-3, shannon-3-pro, shannon-3.1, shannon-3.1-pro Litšoantšo, le litokomane: lifaele tsa PDF, Word, PowerPoint, Excel le mongolo o hlakileng.
shannon-1.6-lite, shannon-1.6-pro Litšoantšo.
shannon-2-lite, shannon-2-pro Litšoantšo, likopong tse boetseng li romela tools kapa response_format.
Kimi-K3-3BIT-REAP, MiniMax-M3-3BIT-REAP, Kimi-K2.6-W4A16-AUTOROUND-REAP, inkling-W4A16-AUTOROUND-REAP, MiMo-V2.5-W8A16 Litšoantšo.
DeepSeek-V4-Pro-0813-3BIT-REAP, GLM-5.2-3BIT-REAP, Nemotron3Ultra-3BIT-REAP, DeepSeek-V4-Flash-0731-W4A16-AUTOROUND-REAP, Laguna-S-2.1-W4A16-AUTOROUND-REAP, MiMo-V2.5-Pro-W8A16, Hy3-W8A16 Mongolo. Kopo e nang le setšoantšo e arajoa ka boemo ba 400.
shannon-coder-1 Mongolo.

GET /v1/models e tlaleha input ea setšoantšo ka id e le capabilities.vision. Li-model le litheko

  • Li-model tsa Shannon li bala lifaele tsa molaetsa oa ho qetela oa user. Behe setšoantšo molaetsing o botsang ka sona.
  • Lelapa la Shannon 3 le boetse le boloka litšoantšo le litokomane tse 'ngoe tsa melaetsa e fetileng ea user ha li ne li romeloa inline, mme le bala palo e phahameng e behiloeng ea litšoantšo ka kopo, ho tloha ho ea pele.
  • Li-model tse hosted tsa open-weight li bala litšoantšo tsa molaetsa o mong le o mong oa moqoqo.

Mebopeho e amohelwang

Faele e romeloa ka tsela e le 'ngoe ho tse peli: ka hare ho kopo e le base64, kapa e le aterese eo API e e fumanang. Faele e 'ngoe le e 'ngoe e tsamaea le kopo; API ha e na endpoint ea upload.

Endpoint Ka kopong (base64) Ka aterese
/v1/chat/completions {"type": "image_url", "image_url": {"url": "data:image/jpeg;base64,…"}} {"type": "image_url", "image_url": {"url": "https://…"}}
/v1/responses {"type": "input_image", "image_url": "data:image/jpeg;base64,…"} {"type": "input_image", "image_url": "https://…"}
/v1/messages {"type": "image", "source": {"type": "base64", "media_type": "image/jpeg", "data": "…"}} {"type": "image", "source": {"type": "url", "url": "https://…"}}

Molaetsa o tšoanang oa user sebopeho se seng le se seng:

{
  "role": "user",
  "content": [
    {"type": "text", "text": "What is in this picture?"},
    {
      "type": "image_url",
      "image_url": {
        "url": "data:image/jpeg;base64,/9j/4AAQSkZJRgABAQAAAQABAAD…"
      }
    }
  ]
}
  • Data URL e na le sebopeho data:<media type>;base64,<data>. Karolo ea ;base64 e hlokahala.
  • Ho /v1/chat/completions le /v1/responses, image_url e ka ba string ka boeona kapa ntho {"url": "…"}.
  • Mofuta oa faele o tsoa ho data URL kapa ho media_type. Ntle le oona, setšoantšo se baloa e le image/png.
  • Lifaele tsa inline li bala ho boholo ba body ea kopo, e ka ba 32 MiB.
  • E amohelwa bakeng sa compatibility: image_url.detail.

Meeli ea lifaele tse hole

Ha u romela aterese, API e fumana faele pele model e sebetsa, kahare ho meeli ena:

Moeli Molao
Boholo Ho fihlela ho 8 MiB ka faele.
Nako Metsotsoana e 20 ea ho araba.
Li-redirect Bonyane ba 5.
Aterese http:// kapa https://, ntle le lebitso la user kapa password aterese-ng, ho host e nang le aterese ea sechaba.
Karabo ea host Boemo ba mofuta oa 200 le body e seng ntle. Mofuta oa faele o baloa ho Content-Type.
E kopiloe e le User-Agent: ShannonBot/1.0 (+https://shannon-ai.com)

Hore na litšoantšo li baloa joang ka tokens

Ho li-model tse hosted tsa open-weight setšoantšo se bala token e le 'ngoe bakeng sa patch e 'ngoe le e 'ngoe ea pixel tsa 28 × 28: bophara bo arotsoe ka 28, bo hlatsoeng holimo, ho atolosoa ka bophahamo bo arotsoeng ka 28, bo hlatsoeng holimo. Sephetho ke karolo ea usage.prompt_tokens mme se lefisoa e le input.

1024 × 1024
1,369 tokens
1920 × 1080
2,691 tokens
512 × 512
361 tokens
  • Setšoantšo se hole se fumanoa pele, kahoo boholo ba sona ba 'nete bo a baloa.
  • Setšoantšo seo boholo ba sona bo sa balehe se bala tokens tse 1,024.
  • Li-endpoint tsa ho bala li sebelisa molao o tšoanang bakeng sa li-data URL. Ha li fumane aterese e hole mme li e bala e le tokens tse 1,024. Ho bala tokens
  • Bakeng sa model ea Shannon, bala hore na kopo e nang le litšoantšo e jele bokae ho usage karabong.

Litokomane le lifaele tse ling

Lelapa la Shannon 3 (shannon-3, shannon-3-pro, shannon-3.1, shannon-3.1-pro) le bala le litokomane. Mongolo oa tokomane oa ntšoa mme oa fuoa model hammoho le molaetsa oa hau.

Tokomane Mofuta oa media
PDF application/pdf
Word application/vnd.openxmlformats-officedocument.wordprocessingml.document
PowerPoint application/vnd.openxmlformats-officedocument.presentationml.presentation
Excel application/vnd.openxmlformats-officedocument.spreadsheetml.sheet

Tokomane ke karolo ea molaetsa oa user, joalo ka setšoantšo:

Endpoint Ka kopong (base64) Ka aterese
/v1/chat/completions, /v1/messages {"type": "document", "source": {"type": "base64", "media_type": "application/pdf", "data": "…"}} {"type": "document", "source": {"type": "url", "url": "https://…"}}
/v1/responses {"type": "input_file", "file_data": "data:application/pdf;base64,…"} {"type": "input_file", "file_url": "https://…"}
{
  "model": "shannon-3",
  "messages": [
    {
      "role": "user",
      "content": [
        {
          "type": "document",
          "source": {
            "type": "base64",
            "media_type": "application/pdf",
            "data": "JVBERi0xLjcKJeLjz9MK…"
          }
        },
        {"type": "text", "text": "Summarise this report in five points."}
      ]
    }
  ]
}
  • Fana ka mofuta oa 'nete oa media oa faele. Ntle le oona, tokomane e baloa e le application/pdf.
  • Faele ea mongolo o hlakileng, e kang CSV, e baloa e le mongolo mme e kenngoa ka tsela e tšoanang.
  • Ha tokomane e sa baloe, kapa faele e le ea mofuta o mong, model e bolelloa joalo mme e ka e bolela karabong ea eona. Boemo bo ntse bo le 200.
  • Ho model e 'ngoe le e 'ngoe karolo ea tokomane e siuoa ho seo model e se balang, mme kopo e arajoa ho latela se setseng.
  • Tokomane e rometsoeng ka aterese e fumanoa tlasa meeli ea lifaele tse hole e ka holimo.

Liphoso

Boemo Mofuta Molaetsa Neng
400 invalid_request_error <id> does not accept image input Setšoantšo se rometsoe ho model e hosted ea open-weight e amohelang mongolo feela. Ha ho letho le nkoang ho balance ea hau.
400 invalid_request_error audio and video input are not supported by any hosted model Karolo ea audio kapa video e rometsoe ho model e hosted ea open-weight.
413 invalid_request_error Failed to buffer the request body: … Body ea kopo, hammoho le lifaele tsa eona tsa inline, e kholo ho feta 32 MiB. Phoso e na le code: "request_too_large".

Model ea Shannon e araba ka boemo ba 200 ho sa tsotellehe hore na kopo e nka lifaele life: karolo eo e sa e balang e siuoa. Diphoso