AI Translator & Summarizer - OpenRouter BYO Key, Bulk URL & Text Processing

Transform any URL or raw text through five AI-powered modes: translate, summarize, rewrite, extract key points, and TL;DR. Bring your own OpenRouter API key to select from GPT-4o, Claude Sonnet, Gemini Flash, Mistral, and other models. URL inputs are automatically fetched and extracted to clean article text before processing. Every dataset row includes token counts and vendor cost in USD.

What does it do?

• Translate — Render text in any language. Supply an ISO 639-1 code (e.g. es) or a language name (e.g. French).
• Summarize — Condense content in formal, casual, technical, marketing, or SEO style.
• Rewrite — Reword content in a chosen style while preserving length and meaning.
• Extract key points — Produce a bulleted list of the core ideas from any content.
• TL;DR — One-paragraph summary regardless of input length.

URL inputs go through a full fetch and article extraction pipeline that strips navigation, ads, and boilerplate before the AI call. Raw text inputs are processed directly.

What data does it return?

Each input item produces one dataset row:

Field	Type	Description
input	object	Original input object {url?, text?}
mode	string	Operation mode: translate, summarize, rewrite, extract-key-points, tldr
model	string	OpenRouter model ID used
output	string	Transformed content
inputTokens	number	Tokens consumed by the prompt
outputTokens	number	Tokens consumed by the completion
costUsd	number	Vendor cost in USD for this record
cached	boolean	True when served from local cache (no LLM call made)
extractedFromUrl	boolean	True when a URL was fetched and extracted to text
sourceLanguage	string	Detected or declared source language
targetLanguage	string	Target language (translate mode only)
status	string	success, timeout, or error
errorMsg	string	Error detail on failure

Sample output: translate mode

{
  "input": { "url": "https://example.com/article" },
  "mode": "translate",
  "model": "openai/gpt-4o-mini",
  "output": "El rápido zorro marrón salta sobre el perro perezoso.",
  "inputTokens": 312,
  "outputTokens": 298,
  "costUsd": 0.00000918,
  "cached": false,
  "extractedFromUrl": true,
  "sourceLanguage": "en",
  "targetLanguage": "es",
  "status": "success"
}

Sample output: extract-key-points mode

{
  "input": { "text": "Long article text here..." },
  "mode": "extract-key-points",
  "model": "openai/gpt-4o-mini",
  "output": "- The study examined 1,200 participants over 5 years\n- Primary finding: sleep duration below 6 hours correlates with 23% higher cardiovascular risk\n- Confounders: age, BMI, and exercise frequency were controlled",
  "costUsd": 0.00001240,
  "status": "success"
}

How to use it

Field	Required	Description
inputs	Yes	Array of {url?, text?} objects. Each item can have a URL, raw text, or both.
openrouterApiKey	Yes	Your OpenRouter API key (sk-or-...). Get one free at openrouter.ai.
mode	No	translate (default), summarize, rewrite, extract-key-points, tldr
targetLanguage	Required for translate	ISO 639-1 code or language name
style	No	formal (default), casual, technical, marketing, seo — applies to summarize and rewrite
model	No	OpenRouter model ID. Default: openai/gpt-4o-mini
maxOutputTokens	No	Cap on response length per input. Default: 1024
useCache	No	Cache results by input hash (7-day TTL). Default: true
maxItems	No	Cap on number of inputs to process. Default: 15

Example: bulk translate URLs to Spanish

{
  "inputs": [
    { "url": "https://example.com/article-1" },
    { "url": "https://example.com/article-2" }
  ],
  "mode": "translate",
  "targetLanguage": "es",
  "model": "openai/gpt-4o-mini",
  "openrouterApiKey": "sk-or-your-key"
}

Example: extract key points from raw text

{
  "inputs": [
    { "text": "Your long document text here..." }
  ],
  "mode": "extract-key-points",
  "model": "openai/gpt-4o-mini",
  "openrouterApiKey": "sk-or-your-key"
}

Model selection

Model	Best for
openai/gpt-4o-mini	General translation and summarization; most tasks at lowest cost
google/gemini-flash-1.5	Fastest option for high-volume batch jobs
mistralai/mistral-small	European language translation
anthropic/claude-3.5-haiku	Nuanced rewrites and tone matching
openai/gpt-4o	Highest quality; recommended for documents over 10,000 words or complex technical/legal text
anthropic/claude-3.5-sonnet	Complex content requiring careful reasoning or multi-step extraction

For documents over 10,000 words or for technical, legal, or domain-specific content where accuracy matters, consider upgrading from gpt-4o-mini to gpt-4o or claude-3.5-sonnet.

Use cases

• Content localization — translate web pages or documents into multiple target languages at batch scale
• Newsletter or article digests — summarize a list of URLs into TL;DR cards for digest emails or dashboards
• Knowledge base preparation — extract key points from long documents before importing into Confluence, Notion, or a RAG pipeline
• SEO content rewriting — rewrite existing articles in SEO style to avoid duplication penalties
• Research synthesis — extract structured bullet points from academic papers or reports for comparison

FAQ

Where are API keys stored? Your OpenRouter key is passed directly to the OpenRouter API for each call. It is not stored by the actor beyond the current run.

What is the caching behavior? When useCache is enabled (default), results are cached using a hash of the input text, mode, target language, model, and style. Cache TTL is 7 days. A cache hit returns the stored output with zero token consumption.

What happens when an input fails? Failed inputs produce a row with status: "error" and an errorMsg. The actor continues processing remaining inputs. Common causes: invalid API key, unreachable URL, or missing targetLanguage when mode is translate.

Output is available in JSON, CSV, and Excel via the Apify dataset export panel.