Built by Metorial, the integration platform for agentic AI.
Retrieve, update, or delete an existing web monitor. Can also list all monitors or fetch events for a specific monitor. Use the monitor ID from the **Create Monitor** tool.
Send messages to Parallel's web-grounded chat completions API and receive AI responses with optional citations. Supports multiple models: **speed** for low-latency responses, and **lite**/**base**/**core** for deeper research with citations. Compatible with the OpenAI chat completions format.
Discover entities (companies, people, products, etc.) matching specific criteria from web data. Uses Parallel's FindAll API to generate candidates, validate them against your conditions, and return matches. This is an asynchronous operation — returns a run ID. Use the **Get FindAll Results** tool to poll status and retrieve matched entities.
Extract and retrieve content from specific URLs. Provide one or more URLs with an optional objective to focus extraction on relevant information. Returns excerpts and/or full content from the target pages.
Check the status of a FindAll entity discovery run and retrieve matched entities when complete. Use the run ID returned by the **Find Entities** tool.
Run a complex, multi-step web research task that autonomously searches, retrieves, reasons, and synthesizes information. Declare what information you need and the API orchestrates querying, ranking, retrieval, reasoning, validation, and synthesis. Returns a run ID that can be used with the **Get Task Run** tool to check status and retrieve results when complete.
Create a web monitor that continuously tracks the web for changes relevant to a query on a configurable schedule. Monitors detect new information as it's published and can deliver results via webhooks.
Check the status of a deep research task run and optionally retrieve its results when complete. Use the run ID returned by the **Deep Research** tool.
Perform AI-optimized web searches using natural language objectives and optional keyword queries. Returns ranked results with excerpts and source URLs. Supports three modes: **one-shot** for comprehensive results, **agentic** for concise token-efficient results in agentic loops, and **fast** for lower-latency results. At least one of `objective` or `searchQueries` must be provided.
Add structured enrichment to a completed or running FindAll entity discovery run. Extracts additional fields from matched entities using a JSON schema you define. Enrichment runs asynchronously — results will be available when polling with the **Get FindAll Results** tool.
Convert a natural language objective into structured FindAll parameters. Parses the objective into an entity type and match conditions that can be used with the **Find Entities** tool. Use this to auto-generate the entityType and matchConditions from a plain-English query.