Skip to main content
The Search API returns ranked, LLM-optimized excerpts from web sources based on natural language objectives or keyword queries. Results are designed to serve directly as model input, enabling faster reasoning and higher-quality completions with minimal post-processing.

​
Key Benefits

  • Context engineering for token efficiency: The API ranks and compresses web results based on reasoning utility rather than human engagement, delivering the most relevant tokens for each agent’s specific objective.
  • Single-hop resolution of complex queries: Where traditional search forces agents to make multiple sequential calls, accumulating latency and costs, Parallel resolves complex multi-topic queries in a single request.
  • Multi-hop efficiency: For deep research workflows requiring multiple reasoning steps, agents using Parallel complete tasks in fewer tool calls while achieving higher accuracy and lower end-to-end latency.

​
Request Fields

Note that at least one of objective or search_queries is required. The remaining fields are optional. See the API Reference for complete parameter specifications and constraints.
FieldTypeNotesExample
objectivestringNatural-language description of the web research goal, including source or freshness guidance and broader context from the task. Maximum 5000 characters.”I want to know when the UN was founded. Prefer UN’s websites.”
search_queriesstring[]Optional search queries to supplement the objective. Maximum 200 characters per query.[“Founding year UN”, “Year of founding United Nations”]
max_resultsintMaximum number of search results to return (1-20).10
max_chars_per_resultintMaximum characters per search result (100-~30,000).6000
source_policySourcePolicyControls specific domains to include or exclude from search results. Use only when source guidance in the objective is insufficient.Source policy example

​
Objective and Search Queries

For best results, provide both objective and search_queries. The objective should include context about your broader task or goal, while search queries ensure specific keywords are prioritized. When writing objectives, be specific about preferred sources, include freshness requirements when relevant, and specify desired content types (e.g., technical documentation, peer-reviewed research, official announcements). Examples of effective objectives with search queries: 
{
  "objective": "I'm helping a client decide whether to lease or buy an EV for their small business in California. Find information about federal and state tax credits, rebates, and how they apply to business vehicle purchases vs leases.",
  "search_queries": ["EV tax credit business", "California EV rebate lease", "federal EV incentive purchase vs lease"]
}

{
  "objective": "I'm preparing Q1 2025 investor materials for a fintech startup. Find recent announcements (past 3 months) from the Federal Reserve and SEC about digital asset regulations and banking partnerships with crypto firms.",
  "search_queries": ["Federal Reserve crypto guidance 2025", "SEC digital asset policy", "bank crypto partnership regulations"]
}

{
  "objective": "I'm designing a machine learning course for graduate students. Find technical documentation and API guides that explain how transformer attention mechanisms work, preferably from official framework documentation like PyTorch or Hugging Face.",
  "search_queries": ["transformer attention mechanism", "PyTorch attention documentation", "Hugging Face transformer guide"]
}

{
  "objective": "I'm writing a literature review on Alzheimer's treatments for a medical journal. Find peer-reviewed research papers and clinical trial results from the past 2 years on amyloid-beta targeted therapies.",
  "search_queries": ["amyloid beta clinical trials", "Alzheimer's treatment research 2023-2025", "monoclonal antibody AD trials"]
}