Accounts and API keys

RSS Retriever uses external services with your own accounts. Enter the keys of the services you want on RSS Retriever › Accounts and save with Update settings. You pay each provider directly, at its own prices.

Integrated services

The first tab lists every built-in service. Fill in only the ones you use.

Accounts integrated services and request limitsClick to enlarge
ServiceFieldsUsed for
OpenRouterAPI key, request limitAny model on OpenRouter: [or_text], [ai_generate] and AI engine names in OpenRouter format, [or_image] and the OpenRouter image generator.
OpenAIAPI key, GPT request limit, Image request limitGPT models ([openai_gpt], openai-... engines), GPT Image ([dalle], OpenAI image generator).
Anthropic ClaudeAPI key, request limitClaude models: [claude], anthropic-... engines.
Google GeminiAPI key, request limitGemini models: [gemini], google-... engines.
xAIAPI key, request limitGrok models: [xai], xai-... engines.
GoAPIAPI key, request limitMidjourney and FLUX through GoAPI.
CometAPIAPI key, request limitMidjourney and FLUX through CometAPI.
Stability AIAPI key, request limitStable Diffusion and Stable Image models.
SpinnerChiefAPI key, developer key, request limitThe SpinnerChief content spinner.
SpinRewriterAPI key, email, request limitThe SpinRewriter content spinner.
WordAiAPI key, email, request limitThe WordAi content spinner.
DeepLAPI key, request limitDeepL translation (Pro and Free API keys).
Google TranslateAPI key, request limitGoogle Cloud Translation.
Yandex TranslateAPI key, request limitYandex Translate (API v1.5 and v2 keys).
PixabayAPI key (free)Pixabay images: the Pixabay image option, [pixabay] and %pixabay[...]%.

When a feed needs a service without a key, the request is skipped and the log says so, for example “Anthropic Claude API key is not set”. If that step is required for the post, the post is not published.

Request limits

Every service has a request limit: a number and a period of a hour, a day or a month. When the limit is reached, RSS Retriever stops calling that service until the next period starts, and the log reports that the limit has been reached. 0 means no limit.

Limits protect your budget from surprises, for example a feed that suddenly publishes 200 items, or a prompt that generates more requests than you expected. Remember that one post can use several requests: one per AI shortcode, one per section of [gpt_article], one per generated image.

Custom AI endpoints

The second tab connects any service with an OpenAI-compatible chat API: DeepSeek, Mistral, Groq, Together, a local server or a self-hosted model. Each endpoint has:

Endpoint ID (required)
A unique name of lowercase letters, numbers and underscores, for example deepseek. You use it in shortcodes and engine names.
Endpoint URL (required)
The full URL of the chat completions API, for example https://api.deepseek.com/chat/completions.
Default model
Optional. The model used when a shortcode does not name one.
API key
Optional. Sent as a Bearer token if the service needs one.

Click Add endpoint; use Edit and Delete in the table to change endpoints. Changes are stored only after you click Update settings.

Use an endpoint with the [custom_ai] shortcode, or anywhere an AI engine is expected as endpoint_id-model:

[custom_ai id="deepseek" model="deepseek-chat" prompt="Summarize: %post_content_notags%"]
[ai_generate engine="deepseek-deepseek-chat" prompt="Summarize: %post_content_notags%"]
Custom AI endpointsClick to enlarge
Keep your keys private

Never post screenshots or logs with API keys on the forum or anywhere else. Anyone with your key can spend your credits.