Gemini generateContent
Menu
Chat

Gemini generateContent

Call models in the native Google Gemini API format.

POST /v1beta/models/{model}:generateContent

Compatible with the Google Gemini API (v1beta), so the official google-genai SDKs and Gemini-native tools work with RuToken. Use :generateContent for a single response and :streamGenerateContent for streaming.

Authentication

x-goog-api-key: sk-xxxx, or the query parameter ?key=sk-xxxx. Authorization: Bearer sk-xxxx also works.

Path parameters

Parameter Type Required Description
model string Yes Model ID, e.g. gemini-2.5-flash.

The action after the colon is generateContent or streamGenerateContent.

Query parameters

Parameter Type Required Description
key string No API key, if not sent in a header.
alt string No sse — return the stream as Server-Sent Events. Use with streamGenerateContent.

Request body

Parameter Type Required Description
contents array Yes Conversation turns. Each item has role (user or model) and parts.
systemInstruction object No System prompt: {"parts": [{"text": "..."}]}.
generationConfig object No Generation settings. See generationConfig below.
tools array No functionDeclarations, googleSearch, codeExecution, urlContext — as supported by the model.
toolConfig object No functionCallingConfig.mode: AUTO, ANY, NONE; optional allowedFunctionNames.
safetySettings array No Items with category (e.g. HARM_CATEGORY_HARASSMENT) and threshold (e.g. BLOCK_ONLY_HIGH, BLOCK_NONE).
cachedContent string No Name of cached content to use as context.

Parts

Part Description
text Plain text.
inlineData {"mimeType": "image/png", "data": "<base64>"} — images, audio, video, PDF.
fileData {"mimeType": "...", "fileUri": "..."} — a file referenced by URI.
functionCall / functionResponse Tool call by the model and the result returned by your code.

generationConfig

Field Type Description
temperature number Sampling temperature, 0–2.
topP number Nucleus sampling.
topK integer Top-K sampling.
maxOutputTokens integer Maximum number of tokens to generate.
candidateCount integer Number of candidates to return.
stopSequences array Up to 5 strings that stop generation.
responseMimeType string text/plain or application/json.
responseSchema object Output schema when responseMimeType is application/json.
responseModalities array e.g. ["TEXT", "IMAGE"] for image-capable models.
thinkingConfig object thinkingBudget (tokens, 0 disables thinking on supported models, -1 dynamic), includeThoughts, or thinkingLevel.
seed integer Seed for more deterministic output.
presencePenalty / frequencyPenalty number Repetition penalties.

Request example

cURL
curl "https://rutoken.ai/v1beta/models/gemini-2.5-flash:generateContent" \
  -H "Content-Type: application/json" \
  -H "x-goog-api-key: $API_KEY" \
  -d '{
    "systemInstruction": {"parts": [{"text": "You are a helpful assistant."}]},
    "contents": [
      {"role": "user", "parts": [{"text": "Explain recursion in one sentence."}]}
    ],
    "generationConfig": {"temperature": 0.7, "maxOutputTokens": 512}
  }'
Python
from google import genai
from google.genai import types

client = genai.Client(api_key="sk-xxxx", http_options={"base_url": "https://rutoken.ai"})
resp = client.models.generate_content(
    model="gemini-2.5-flash",
    contents="Explain recursion in one sentence.",
    config=types.GenerateContentConfig(
        system_instruction="You are a helpful assistant.",
        temperature=0.7,
        max_output_tokens=512,
    ),
)
print(resp.text)
Node.js
import { GoogleGenAI } from "@google/genai";

const ai = new GoogleGenAI({ apiKey: "sk-xxxx", httpOptions: { baseUrl: "https://rutoken.ai" } });
const resp = await ai.models.generateContent({
  model: "gemini-2.5-flash",
  contents: "Explain recursion in one sentence.",
  config: {
    systemInstruction: "You are a helpful assistant.",
    temperature: 0.7,
    maxOutputTokens: 512,
  },
});
console.log(resp.text);

Response

json
{
  "candidates": [
    {
      "content": {
        "role": "model",
        "parts": [
          {"text": "Recursion is when a function calls itself to solve smaller versions of the same problem."}
        ]
      },
      "finishReason": "STOP",
      "index": 0
    }
  ],
  "usageMetadata": {
    "promptTokenCount": 12,
    "candidatesTokenCount": 18,
    "thoughtsTokenCount": 0,
    "totalTokenCount": 30
  },
  "modelVersion": "gemini-2.5-flash"
}
Field Type Description
candidates[].content.parts array Generated parts: text, inlineData (e.g. images), functionCall. Parts with thought: true contain thinking summaries.
candidates[].finishReason string STOP, MAX_TOKENS, SAFETY, RECITATION, etc.
usageMetadata object Token counts: prompt, candidates, thoughts and total.

Streaming

bash
curl "https://rutoken.ai/v1beta/models/gemini-2.5-flash:streamGenerateContent?alt=sse" \
  -H "Content-Type: application/json" \
  -H "x-goog-api-key: $API_KEY" \
  -N \
  -d '{"contents": [{"role": "user", "parts": [{"text": "Write a short poem."}]}]}'

Each data: event is a GenerateContentResponse with the next fragment. See Streaming.

Notes

  • Gemini models can also be called through Chat Completions in OpenAI format.
  • Errors on this endpoint are returned in the OpenAI error format. See Errors.
  • List models in Gemini format with GET /v1beta/models. See List models.
← Previous
Claude Messages
Next →
Completions (legacy)