Research Agent

Showing 25 of 26 from the live probe.

• Research

  Run a deep-web research query (costs 1 credit). Choosing a model + writing a schema: call list_models() for the full guide. Quick version: openai/gpt-5-nano = cheapest to run + fast at top-tier quality; google/gemini-3.1-flash-lite = balanced default; google/gemini-2.5-flash (2 credits) = highest measured quality; 3.5-flash (6 credits) is premium-cost with no quality edge — reserve it. Raise `depth` before the model tier. For structured output keep the schema flat, describe

• List Agents

  List all your configured research agents with their IDs, models, and input variables.

• Get Agent

  Get full details of a specific agent by ID. Args: agent_id: UUID of the agent.

• Create Agent

  Create a new research agent. Use {{variable_name}} placeholders in the prompt_template that match input_variables. Before you pick a model or write output_schema, call list_models() for the model-fit + schema-authoring guide. Short version: openai/gpt-5-nano (1 credit) is cheapest to run and fast at top-tier quality; google/gemini-3.1-flash-lite (1) is the balanced default; google/gemini-2.5-flash (2) is the highest-quality option. Schemas: flat + shallow, describe ever

• Create Agent From Template

  Create a new agent from a pre-built template (use list_templates first). Args: template_key: The template key (e.g. "icp_check", "company_enrichment").

• Update Agent

  Update an existing agent's configuration. Only provided fields are changed. Args: agent_id: UUID of the agent to update. name: New name. description: New description. model: New model ID (included: openai/gpt-5-nano cheapest 1 credit, google/gemini-3.1-flash-lite default 1, google/gemini-2.5-flash highest quality 2, 3.5-flash premium 6). See list_models(). prompt_template: New prompt template with {{variable}} placeholders. Include a no-in

• Delete Agent

  Permanently delete an agent. Args: agent_id: UUID of the agent to delete.

• Run Agent

  Execute a configured research agent with input variables (costs 1 credit). Args: agent_id: UUID of the agent to run (get from list_agents). input_variables: JSON object with the agent's required variables. Example: '{"company_name": "Acme", "website": "acme.com"}' max_iterations: Max agent iterations 1-50 (default 25). Overrides depth preset if set. depth: Research depth — "quick" (fast, 1-3 tools), "standard" (default), or "deep" (thorough, 8-15

• List Agent Runs

  List past runs for an agent with their results. Args: agent_id: UUID of the agent. limit: Max number of runs to return (default 20, max 200).

• List Templates

  List available pre-built agent templates you can create agents from.

• Check Credits

  Check your current credit balance and recent transactions.

• Get Usage

  Get your usage statistics for the last N days. Args: days: Number of days to look back (default 30).

• Get Me

  Get the current user's identity (id, email, display_name, created_at).

• Bulk Run Agent

  Submit a bulk run job to execute an agent on many inputs at once. Costs 1 credit per item. Items run concurrently (up to 5 at a time). Returns a job_id you can poll for progress. Args: agent_id: UUID of the agent to run. items: JSON array of objects, each containing the agent's input variables. Example: '[{"company_name": "Acme", "website": "acme.com"}, {"company_name": "Globex", "website": "globex.com"}]' max_iterations: Max agent iterations per

• Get Bulk Run Status

  Check the progress of a bulk run job. Args: agent_id: UUID of the agent. job_id: UUID of the bulk job (from bulk_run_agent).

• Get Bulk Run Results

  Get results from a bulk run job (supports pagination). Args: agent_id: UUID of the agent. job_id: UUID of the bulk job. offset: Number of items to skip (default 0). limit: Max items to return (default 50).

• Cancel Bulk Run

  Cancel a running bulk job. Items already completed are kept. Args: agent_id: UUID of the agent. job_id: UUID of the bulk job to cancel.

• List Bulk Runs

  List all bulk run jobs for an agent. Args: agent_id: UUID of the agent.

• Search Web

  Search the web for a query and return results (costs 1 credit). A lightweight web search — no deep analysis, just search results. Args: query: The search query.

• Scrape Page

  Scrape and summarize the content of a web page (costs 1 credit). Args: url: The full URL to scrape (e.g. "https://example.com/page").

• Enrich Company

  One-shot company enrichment — returns structured company data (costs 1 credit). Args: company_name: Name of the company to enrich. domain: Optional company domain/website for more accurate results.

• Find Buying Signals

  Find recent buying signals for a company — funding, hires, launches, expansions (costs 1 credit). Args: company_name: Name of the company to research.

• Score Icp Fit

  Score how well a company fits your Ideal Customer Profile (costs 1 credit). Args: company_name: Name of the company to evaluate. icp_description: Description of your ideal customer profile (e.g. "B2B SaaS companies, 50-500 employees, Series A-C, in fintech or healthtech").

• List Models

  List available models WITH guidance on which model fits which task and how to set up a prompt + schema correctly. Read this before choosing a model or writing an output schema for research()/create_agent().

• Get Plans

  List available billing plans and their prices.