---
title: "Olostep integration for LuumenAI"
url: "https://www.luumen.ai/integrations/olostep/"
entity: SoftwareIntegration
category: "Databases & data platforms"
audience: "SRE, DevOps, platform engineering, IT operations"
---

# Olostep integration for LuumenAI

Scrape, crawl, search, and research the web from the same terminal session

## What LuumenAI can do with Olostep

LuumenAI connects to Olostep with an API token and exposes 9 tools (4 read-only · 5 write). Reads answer immediately; writes are shown as a plan and require approval by default, configurable per tool by an administrator.

## Key facts

- Authentication: API token
- Tools available: 9 tools (4 read-only · 5 write)
- Plan: available on all plans
- Category: Databases & data platforms

## About Olostep with LuumenAI

Olostep gives Luumen a way to read the public web during an investigation. Pull a vendor status page, map a documentation site, search for a known error string, or hand over a bounded research task and get an answer with sources. Credit balance and stored results are readable on their own, so you can check cost and re-read past work before spending anything new.

## Example tools

Read-only (4 in total):

- `OLOSTEP_GET_CREDIT_BALANCE` — Return the connected team's remaining Olostep credits, credit lots, active subscription, and whether usage is enabled without consuming web-processing credits.
- `OLOSTEP_GET_OPERATION_RESULT` — Retrieve a stored scrape, search, map, answer, or crawl by its operation ID without repeating or rebilling the original work.
- `OLOSTEP_LIST_CRAWL_PAGES` — Return one page of URLs and retrieve IDs processed by an asynchronous crawl, with a continuation cursor for additional pages.
- `OLOSTEP_RETRIEVE_CONTENT` — Retrieve stored page content by retrieve ID in selected live-supported formats, including hosted URLs for oversized output.

Write and destructive (5 in total, approval required by default):

- `OLOSTEP_ANSWER_QUESTION` — Synchronously research a bounded task by searching and browsing the web, then return the completed answer and sources when available.
- `OLOSTEP_MAP_WEBSITE` — Synchronously discover and optionally rank URLs on a website, returning one response page and a continuation cursor.
- `OLOSTEP_SCRAPE_PAGE` — Synchronously scrape one web page and return requested content, links, metadata, a retrieve ID, and billing usage.
- `OLOSTEP_SEARCH_WEB` — Synchronously search the web for relevant deduplicated links.
- `OLOSTEP_START_CRAWL` — Start an asynchronous, usage-metered website crawl and immediately return its crawl ID and initial status.

## Frequently asked questions

### How does LuumenAI connect to Olostep?

Authorize once with API token. Luumen lists the scopes each action needs before you approve the connection, and credentials never appear in the chat.

### Can LuumenAI change things in Olostep on its own?

Read actions answer immediately. Anything that writes is shown as a plan and requires approval by default. Administrators configure that per tool.

### Is there an audit trail?

Every call to Olostep — read or write, approved or declined — is recorded with the actor, the input, and the result.

Full page: https://www.luumen.ai/integrations/olostep/