AI AGENT STORE / SERVICES

SEARCH & RESEARCH · FOR AI AGENTS

Valyu

Search web and specialist research sources through one API.

Retrieve source material from web, academic, financial and other specialist indexes, or start a cited research task through Valyu’s API and hosted MCP tools.

Your integration checklist

What you need from your owner
Create a Valyu account, obtain an API key or authorize MCP sign-in, and select a plan with access to the datasets your task needs.
Authentication
Direct API calls use x-api-key. Hosted MCP supports sign-in or an API key; header authentication avoids placing a secret in a logged connection URL.
Your first task
Search a narrowly defined question, retain the returned source URLs and check one important result against its source.
Billing & payment
Usage consumes credits at source- and endpoint-specific rates. Dataset access depends on the plan; new requests pause at zero balance unless credits are replenished, including through enabled auto top-up.
Agent payment support
Not verified. Programmatic usage does not establish autonomous checkout support.
Operational limits
Dataset coverage and freshness vary. A ranked result or generated report still needs evidence review. Inspect the current request schema and budget units before setting price controls.

ORIGINAL AI AGENT STORE INTEGRATION PLAYBOOK

Keep source coverage visible throughout the research task

Editorial recommendation · Reviewed 2026-10-17

When to choose this service

Choose Valyu when the task needs current web evidence or specialist sources available through its indexes. Confirm dataset access before committing to a research workflow.

Maintained indexes, external source access and current retrieved content are resources outside the agent’s reasoning. A search prompt cannot recreate inaccessible or newly published evidence.

Suggested workflow

  1. Define the question, time range and required source types. Inspect dataset availability and set a bounded result count and credit budget.
  2. Retrieve a small sample, keeping source identity and dates with each claim. Compare a decisive result with the underlying source before expanding.
  3. For a longer research task, retain its task ID and check the existing task until completion. Return citations and unresolved evidence gaps with the answer.

Acceptance checks before real work

  • Distinguish source publication dates from the time of retrieval.
  • Check that an absent result is not caused by a plan restriction or overly narrow filter.
  • Verify that citations support the specific claims attributed to them.

If the workflow fails

If a source is unavailable, report the gap and seek another authorized source. Resume an existing research task by ID after a timeout instead of starting a duplicate paid task.

Provider documentation

Explore services for adjacent needs

Continue building your toolkit