Parallel Monitor integration | LangChain 日本語

Parallel is a real-time web search and content extraction platform built for LLMs and AI applications.

ParallelMonitor calls Parallel's Monitor API to schedule a query on a recurring cadence (1 hour to 30 days) and notify you, via webhook or polling, when relevant new content shows up. Useful for ongoing competitive intelligence, regulatory tracking, or topic monitoring.

Overview

Integration details

Class	Package	Serializable	JS support	Package latest
`ParallelMonitor`	`langchain-parallel`	❌	❌

Setup

The integration lives in the langchain-parallel package.

pip

pip install -U langchain-parallel

uv add langchain-parallel

Credentials

Head to Parallel to sign up and generate an API key. Set PARALLEL_API_KEY in your environment:

import getpass
import os

if not os.environ.get("PARALLEL_API_KEY"):
    os.environ["PARALLEL_API_KEY"] = getpass.getpass("Parallel API key:\n")

Instantiation

from langchain_parallel import ParallelMonitor

client = ParallelMonitor()

Create a monitor

frequency is <n><unit> where unit is h, d, or w, and the resulting duration sits in [1h, 30d] (e.g. "1h", "6h", "3d", "1w", "2w").

from langchain_parallel import MonitorWebhook

monitor = client.create(
    query="Latest peer-reviewed papers on net-energy-gain fusion",
    frequency="1d",
    metadata={"label": "fusion-watch"},
    # Optional: receive new events on a webhook endpoint.
    # webhook=MonitorWebhook(
    #     url="https://example.com/parallel/webhook",
    #     event_types=["monitor.event.detected"],
    # ),
)

print(monitor["monitor_id"], "—", monitor["status"])

The valid event_types are monitor.event.detected, monitor.execution.completed, and monitor.execution.failed. Omit the field to receive all three.

Retrieve, list, delete

got = client.retrieve(monitor["monitor_id"])

monitors = client.list(limit=20)

deleted = client.delete(monitor["monitor_id"])

List events

list_events() returns a flat {"events": [...]} list. Newly-created monitors typically have none until the first scheduled execution; the call shape works either way.

events = client.list_events(monitor["monitor_id"], lookback_period="7d")
for ev in events.get("events", []):
    print(ev["type"], ev.get("event_date"))

To inspect a single event group:

group = client.get_event_group(monitor["monitor_id"], event_group_id="...")

Simulate an event

Useful when wiring up a webhook handler. simulate_event() posts a synthetic event of the requested type (default: monitor.event.detected) to the monitor's webhook.

Note

simulate_event() requires a webhook configured on the monitor. Pass webhook=MonitorWebhook(...) when calling create() first; otherwise the API returns 400 — Webhook not configured for this monitor.

client.simulate_event(monitor["monitor_id"], event_type="monitor.event.detected")

Async

Every CRUD and event method has an async counterpart prefixed with a:

monitor = await client.acreate(query="...", frequency="6h")
events = await client.alist_events(monitor["monitor_id"], lookback_period="7d")
await client.adelete(monitor["monitor_id"])

Webhook signature verification

Webhook deliveries are signed using the Standard Webhooks scheme (HMAC-SHA256 over <webhook-id>.<webhook-timestamp>.<body>, base64-encoded, v1,<sig> format with replay protection). Verify with verify_webhook:

from langchain_parallel import verify_webhook

ok = verify_webhook(
    body,
    webhook_id=request.headers["webhook-id"],
    webhook_timestamp=request.headers["webhook-timestamp"],
    webhook_signature=request.headers["webhook-signature"],
    secret=os.environ["PARALLEL_WEBHOOK_SECRET"],
)

The signing secret is configured at the org webhook-endpoint level in the Parallel dashboard, not per-monitor. See webhook setup for the full delivery contract.

Response format

create() and retrieve() return a monitor dict. list_events() returns a flat events envelope.

# create() / retrieve()
{
    "monitor_id": "mon_abc",
    "status": "active",
    "query": "...",
    "frequency": "1d",
    "metadata": {...},
    "webhook": {"url": "...", "event_types": [...]},
    "created_at": "2026-04-15T10:30:00Z",
    # ... additional API-side fields
}

# list_events()
{
    "events": [
        {
            "type": "event",            # or "completion" / "error"
            "event_date": "2026-04-15T10:30:00Z",
            "event_group_id": "...",
            # ... event-type-specific fields
        },
    ],
}

API reference

For detailed documentation, head to the ParallelMonitor API reference or the Monitor API guides.