Guides

Search, Agent, Workflow or Monitor: Which One Fits the Job?

A decision guide to four social media research tools in Superlurk: a search for a lookup, the agent for a question, a workflow for a report, a monitor over time.

Use a search for one lookup, the agent for an open question, a workflow for a report you will want again, and a monitor for the same search repeated over time. Superlurk puts these four social media research tools in one app, on one credit balance. This guide shows what each costs and returns, and which one fits a given job.

The short version

  • Search: you know the network and the query.
  • Agent: you have a question and do not know which searches answer it.
  • Workflow: you need a structured report, and probably the same kind again later.
  • Monitor: you want the same search run on a schedule and every run kept.

What each one costs and returns

Cost and output of the four tools
What it costsWhat you get back
SearchThe exact price is shown before it runs. Most searches cost 1 credit, and a failed search is free.A list of results, saved to your private Results and exportable as CSV or JSON.
Agent1 credit per Quick reply or 5 per Deep reply, inside a budget you set for the chat (25 credits by default).A written answer with numbered citations that open the cited posts and comments.
WorkflowThe searches that actually run plus an analysis fee of 5 credits (standard) or 25 credits (deep). The ceiling is shown before you run.A report where every claim links to the post behind it, kept for 30 days.
MonitorThe most it can cost is shown before you save it. It pauses itself when credits run out.Every run saved to Results. No alerts, emails or digests.

Search: one lookup

Pick a network and a search, then enter a keyword, #hashtag, @handle or link. Use it when you can already phrase what you want as a query. Superlurk covers 25 networks; the guide to searching all social media from one place shows how to take one question across several of them.

Agent: an open question

The agent is a chat that looks things up as you talk. It runs 1-credit searches on its own inside the chat's budget, and anything more expensive or multi-step comes back as a plan card that waits for you to press Run. Use it when the next search depends on what the last one found. More in what a social search agent is.

Workflow: a report you will want again

A workflow is a fixed chain: search, pick the top items, follow-up searches, analysis, report. Pick a ready-made workflow, fill in a short brief and read “Run · up to N credits” before you run it. Use it when the output has a known shape, such as a brand pulse, a competitor comparison or a creator scorecard. Workflows do not repeat on their own. The workflow guide walks through all 13, and the directory lists them.

Monitor: the same search over time

A monitor re-runs a saved search every 24 hours, every 6 hours or every hour and saves every run to Results. Monitors are on paid plans: Starter allows up to 3, checked daily; Pro up to 15, as often as every 6 hours; Max up to 50, as often as every hour. New-post alerts are not available, so plan a routine for opening the latest run. Setup advice is in the monitoring guide.

Six jobs, matched to a tool

Scenarios and the tool that fits
UseWhy
Looking up one account from its handleSearchOne lookup with a known network and handle.
“Why are people returning this product?”AgentThe agent chooses the searches and cites what it finds.
A monthly read on what people say about your brandWorkflow: Brand pulseA known report shape. You run it each month yourself.
Choosing between four creators for a campaignWorkflow: Creator vetting scorecardThe same checks for each creator, from profile links.
Following a product name through launch weekMonitorThe same search on a schedule, with every run saved to Results.
Making sense of results you already collectedWorkflow: Analyse a saved resultIt runs no searches and returns themes, tone and the most telling items.

How they chain together

  1. Explore with the agent. Ask the question, approve a plan and read the cited answer.
  2. Save the plan as a workflow. A plan you ran can be saved as a workflow, so the same steps are ready for the next brief.
  3. Monitor the query that matters. When one search deserves repeating, save it as a monitor. Each run is saved to Results.
  4. Analyse what the monitor collected. Later, point Analyse a saved result at one of those runs for its themes, tone and most telling items.

What all four share

  • You see the price, the ceiling or the budget before credits are spent.
  • A failed search is free.
  • Only public content can be searched.
  • One credit balance. Plans are on the pricing page.

If you would rather work from an AI assistant you already use, Superlurk's MCP server gives your AI assistant the same searches; see integrations. That is separate from the agent, which is built into Superlurk.

Frequently asked questions

What is the difference between the agent and a workflow?

The agent is a chat for an open question: it chooses the searches as the conversation goes. A workflow is a fixed chain of steps that produces the same kind of report each time from a short brief. A plan you ran with the agent can be saved as a workflow.

Do agent replies and workflows need separate credits?

No. Agent replies and workflows use the same credits as searches, from one balance.

Which tool sends alerts about new posts?

None. New-post alerts are not available. A monitor re-runs a saved search on a schedule and saves every run to Results, where you open the latest run yourself.

How this guide was written. Written by the Superlurk team on 12 November 2026 from the app's current behaviour. Scenarios are illustrative, not measured customer results.

The Superlurk TeamSocial research

We build Superlurk for public-data research across 25 supported social networks, with the price shown before every search. We write practical guides to search methods, source verification and AI-assisted research.

Start lurking for free.

Every network, no card needed. See the price first, and pay nothing if a search fails.