Generate Content


POST https://api.fenno.ai/v1beta/models/{model}:generateContent
x-goog-api-key: <FENNO_API_KEY>
Content-Type: application/json

Use the selected Gemini-compatible model ID. In a raw URL, add the models/ path segment only once; a discovery resource name may already include it.

Request fields

Field	Type	Meaning
`contents`	array, required	Conversation content, usually `role` plus `parts`
`contents[].parts[].text`	string	Text input
`contents[].parts[].inlineData`	object	Supported inline binary input with `mimeType` and base64 `data`
`systemInstruction`	object, optional	Model-supported system instruction parts
`generationConfig`	object, optional	Supported output length, sampling and other generation controls
`tools`, `toolConfig`	array / object, optional	Only tool definitions and behavior supported by the model

Model support is required for multimodal inputs and optional settings. Do not assume a Google-hosted file URI is usable through every gateway.

JavaScript SDK example

Install @google/genai; set FENNO_API_KEY and FENNO_MODEL in a server-side Node.js environment:


import { GoogleGenAI } from '@google/genai'
 
const ai = new GoogleGenAI({
  apiKey: process.env.FENNO_API_KEY,
  httpOptions: { baseUrl: 'https://api.fenno.ai' }
})
 
const response = await ai.models.generateContent({
  model: process.env.FENNO_MODEL!,
  contents: 'Explain what an API gateway does.'
})
console.log(response.text)

For a minimal raw JSON body:


{"contents": [{"role": "user", "parts": [{"text": "Explain what an API gateway does."}]}]}

Response handling

candidates[] contains generated content and finish information; text lives in content.parts[]. The SDK response.text helper is convenient for simple text, but raw clients should inspect the parts rather than assuming one string.

Check candidate finishReason, prompt safety feedback and usage metadata when returned. A blocked prompt or output limit may yield no complete answer; an empty text value alone does not identify the cause. Function calls are typed parts, not text to execute.

Streaming

For REST SSE, use POST /v1beta/models/{model}:streamGenerateContent?alt=sse. For the SDK, replace generateContent with:


const stream = await ai.models.generateContentStream({
  model: process.env.FENNO_MODEL!,
  contents: 'Explain what an API gateway does.'
})
for await (const chunk of stream) {
  if (chunk.text) process.stdout.write(chunk.text)
}

Inspect completion/safety states in the chunks as well; this compact loop only illustrates text rendering. Do not restart a partially delivered stream without telling the caller.

See streaming, vision and error formats.