Skip to main content

​
API Parameters

This guide covers all available parameters for the LLMLayer search API.

​
Required Parameters

​
query
string
required
The search query or question you want to ask. Be as specific as possible for better results.Example: "What are the latest developments in quantum computing?"
​
model
string
required
The LLM model to use for processing the search results.Supported models : -"openai/gpt-5" -"openai/gpt-5-mini" -"openai/gpt-5-nano"
  • "openai/o3"
  • "openai/o3-mini"
  • "openai/o4-mini"
  • "openai/gpt-4.1"
  • "openai/gpt-4.1-mini"
  • "openai/gpt-4o"
  • "openai/gpt-4o-mini"
  • "groq/openai-gpt-oss-120b"
  • "groq/openai-gpt-oss-20b"
  • "groq/kimi-k2"
  • "groq/qwen3-32b"
  • "groq/llama-3.3-70b-versatile"
  • "groq/deepseek-r1-distill-llama-70b"
  • "groq/llama-4-maverick-17b-128e-instruct"
  • "deepseek/deepseek-chat"
  • "deepseek/deepseek-reasoner"
  • "anthropic/claude-sonnet-4"

​
Optional Parameters

​
return_sources
boolean
default:"false"
Include source URLs and metadata in the response. Useful for citations and verification.
​
return_images
boolean
default:"false"
  • Include relevant images found during the search (when available).
  • if set to true, it’s will cost an additional 0.001 $.
​
citations
boolean
default:"false"
Add inline citations [1] in the response text that reference the sources.
​
max_tokens
integer
default:"1500"
Maximum number of tokens in the LLM response. (limit is your model max tokens)
​
temperature
float
default:"0.7"
Controls randomness in the response. Range: 0.0-1.0.
  • Lower values (0.0-0.3): More focused and deterministic
  • Higher values (0.7-1.0): More creative and varied
​
search_type
string
default:"general"
Type of search to perform.Options:
  • "general": Web-wide search
  • "news": Recent news articles only
​
date_filter
string
default:"anytime"
Filter results by recency.Options: "anytime", "hour", "day", "week", "month", "year"
​
location
string
default:"us"
Country code for localized search results. See supported locations below.
​
response_language
string
default:"auto"
Language for the response. "auto" detects from query. See supported languages below.
​
search_context_size
string
default:"medium"
Size of the search context to use for the LLM.Options:
  • "low": Small context, faster but less comprehensive
  • "medium": Balanced context size (default)
  • "high": Large context, more comprehensive but slower and more expensive in input tokens
​
answer_type
string
default:"markdown"
Format of the response.Options:
  • "markdown": Formatted text with headers, lists, etc.
  • "html": HTML formatted response
  • "json": Structured JSON (requires json_schema)
​
json_schema
string
Required when answer_type="json". Defines the expected JSON structure.Example:
{
    "films":[
        {
            "title": {"type": "string"},
            "release_year": {"type": "integer"},
            "director": {"type": "string"},
            "cast": {
            "type": "array",
            "items": {"type": "string"}
            }
        }
    ]
}
​
domain_filter
array
Restrict search to specific domains. Useful for trusted sources.
  • Example: ["pubmed.gov", "nature.com", "arxiv.org"]
  • ["pubmed.com", "nature.com"] to include domains and add "-" to exclude domains, e.g. ["pubmed.com", "-nature.com"]
  • Never use http:// or https:// prefixes, just the domain name.
​
max_queries
integer
default:"1"
  • Number of search queries LLMLayer should generate internally. Range: 1-5.
  • Higher values may find more comprehensive results.
  • each search query will cost 0.004 $.
​
system_prompt
string
Override the default system prompt. Use this for specialized behavior.Example: "You are a medical research assistant. Focus on peer-reviewed sources."
​
provider_key
string
Your API key from your selected model provider.
  • this is an option we have added to allow you to use your own keys and be charged directly by the model provider.
  • sometimes we hit Rate Limits ( specially with Anthropic and Groq ), so you can use your own keys to avoid this.
  • Example: "sk-..." for OpenAI, "sk-ant-..." for Anthropic
  • we support models from 4 providers (OpenAI, Anthropic, DeepSeek, Groq) and you can use your own keys for all of them.

​
Supported Locations

The location parameter accepts country codes for localized results:

🇺🇸 us

United States

🇨🇦 ca

Canada

🇬🇧 uk

United Kingdom

🇲🇽 mx

Mexico

🇪🇸 es

Spain

🇩🇪 de

Germany

🇫🇷 fr

France

🇵🇹 pt

Portugal

🇧🇪 be

Belgium

🇳🇱 nl

Netherlands

🇨🇭 ch

Switzerland

🇳🇴 no

Norway

🇸🇪 se

Sweden

🇦🇹 at

Austria

🇩🇰 dk

Denmark

🇫🇮 fi

Finland

🇹🇷 tr

Turkey

🇮🇹 it

Italy

🇵🇱 pl

Poland

🇷🇺 ru

Russia

🇿🇦 za

South Africa

🇦🇪 ae

UAE

🇸🇦 sa

Saudi Arabia

🇦🇷 ar

Argentina

🇧🇷 br

Brazil

🇦🇺 au

Australia

🇨🇳 cn

China

🇰🇷 kr

Korea

🇯🇵 jp

Japan

🇮🇳 in

India

🇵🇸 ps

Palestine

🇰🇼 kw

Kuwait

🇴🇲 om

Oman

🇶🇦 qa

Qatar

🇮🇱 il

Israel

🇲🇦 ma

Morocco

🇪🇬 eg

Egypt

🇮🇷 ir

Iran

🇱🇾 ly

Libya

🇾🇪 ye

Yemen

🇮🇩 id

Indonesia

🇵🇰 pk

Pakistan

🇧🇩 bd

Bangladesh

🇲🇾 my

Malaysia

🇵🇭 ph

Philippines

🇹🇭 th

Thailand

🇻🇳 vn

Vietnam

​
Supported Languages

The response_language parameter controls the output language:

auto

Auto-detect (default)

en

English

fr

French

es

Spanish

de

German

it

Italian

pt

Portuguese

nl

Dutch

ja

Japanese

ko

Korean

zh

Chinese

ar

Arabic

ru

Russian

tr

Turkish

hi

Hindi