Web research APIs for AI: search, extract, multi-hop research tasks, entity discovery (FindAll), and web monitoring.
Search the Web
Search the Web. Search the web with natural-language objectives. Returns LLM-optimized excerpts with citations
Extract Web Content
Extract Web Content. Extract clean, citation-aware content from specific web URLs. Up to 20 URLs per request
Get FindAll Result
Get FindAll Result. Return the current snapshot of matched candidates for a FindAll run, including any enrichment fields
Create Task Run
Create Task Run. Initiate a Parallel research task. Returns immediately with a queued run; use Get Task Run Result to wait for completion
Get Task Run Status
Get Task Run Status. Retrieve the current status of a task run by its ID
Get Task Run Result
Get Task Run Result. Retrieve a task run result by run ID. Blocks until the run completes (or until the timeout is reached)
Create FindAll Run
Create FindAll Run. Discover and verify entities matching plain-language criteria. Match conditions can be supplied directly or auto-generated from the objective
Chat Completion
Chat Completion. Get a chat completion using a Parallel processor. OpenAI-compatible interface; processor names like 'speed', 'base', 'core', 'pro', 'ultra' map to Parallel tiers
Custom API Call
Custom API Call. Make a custom API call to a specific endpoint
You connect with: api key.
Something happens
An Agent or workflow decides what to do
A person approves, if it matters
Parallel: create task run
Connecting Parallel does not give every Agent unrestricted access to it. Each Agent needs the capability granted, the identity it acts as must be allowed, and consequential actions can wait for a person to approve them.
How Nerve controls actionsStart with one conversation your team handles every day. Give the Agent what it needs to know, connect the tools it should use, and decide where a person steps in.