Watch Plu in Action
Paste your URL and anchor text. Our AI reads your product, scans your partner blogs, and writes the exact paragraph edit, placed where it reads naturally, not where it's easy.
Full pipeline, unedited: scan, score, write, verify.
Built for the people who ship links.
SEO agencies
Running link placement across dozens of client inventories, without a spreadsheet and a headcount for every account.
Link building agencies
Managing partner blog relationships at scale, with full visibility into every placement that goes out under the brand.
Affiliate & partnerships teams
Needing natural-reading placements fast, across every partner in the network, not just the easy ones.
Manual link-building is a spreadsheet nightmare.
of blog posts read by hand, per client, per month.
on which pages actually fit, no real relevance scoring.
Insertions get bolted on instead of written into the piece.
Too slow to scale, or flagged for looking bolted on.
From URL to natural placement in four steps.
Give it a target
Drop in your product URL and anchor text.
It reads your product
AI builds a full profile, audience, category, topics, in seconds.
It scans your inventory
Every blog on your partner domains gets mapped, filtered, and scored for real topical relevance, not keyword stuffing.
It writes the edit
Top 3 placements come back with the exact before/after paragraph, anchor embedded naturally in the body, never the intro, never a bullet list.
One pipeline, twelve specialist agents.
Plu Agent doesn't hand this to one model. Every step is owned by an agent with a single responsibility, so a bad pick, a dead page, or a policy violation gets caught before it reaches you.
Sourcing Agents
Scout
Finds every blog that fits your target, ranked best-first.
Sentinel
Throws out fake and mismatched URLs before you ever see them.
Harvester
Live-scrapes each page's real content, never cached, never twice.
Auditor
Drops dead or already-linked pages, promotes the next best pick.
Placement Agents
Weaver
Reads the full article and writes one contextual insertion, or skips.
Warden
Validates the edit reads natural and every existing link stays intact.
Operations Agents
Quartermaster
Keeps every run inside your monthly plan, automatically.
Foreman
Runs large batches reliably, picking up right where it stopped.
Archivist
Keeps run history and a full audit trail of every placement.
Outreach Agents
Prospector
Finds the right contact when a domain has none on file.
Scribe
Writes the outreach email, pitching your actual placement.
Courier
Sends from your Gmail, tracks the thread and every reply.
Why Plu beats generic AI models for building backlinks.
ChatGPT, Claude, Perplexity, and Gemini weren't built for this. Plu was.
Sourcing from memory, not the live web
A generic model leans on what it was trained on, not a real-time crawl. It doesn't know which blogs are live, updated, or even still online today.
Hallucinated links
Ask a generic model for backlink opportunities and it will confidently hand you URLs that don't exist. Plu verifies every hostname and page before it's ever shown to you.
One pass, not a back-and-forth
Finding and validating a real placement in a chat tool takes round after round of prompting. Plu runs the entire pipeline in a single, unattended pass.
Real crawling, not blocked access
Many blogs sit behind robots.txt rules or paywalls that AI browsing tools respect and skip. Plu indexes and reads the actual page, not a search snippet.
| What matters | Plu Agent | Generic AI models |
|---|---|---|
| Source of data |
✅Live crawl of your partner blogs, every run.
|
❌Training memory, stale, incomplete, or outdated.
|
| Link accuracy |
✅Every URL verified against the live hostname before it
ships.
|
❌Confidently hallucinates URLs that don't exist.
|
| Effort to get a result |
✅One run, fully automated end-to-end.
|
❌Multiple rounds of prompting and manual checking.
|
| Page access |
✅Reads the real, live page, including sites AI browsing
tools can't reach.
|
❌Blocked by robots.txt, paywalls, and JS-only content.
|
| Scale |
✅Dozens of domains, batched and unattended.
|
❌One conversation, one domain, at a time.
|
Rigor on every row.
Precision placements
Relevance-scored against your product, matched to the right content type, and inserted mid-body, never an intro, list item, or bolted-on line. Up to 3 picks per domain, shown before/after with the change highlighted.
Scales across every domain
Every partnered domain runs concurrently, one failure never stalls the rest. Upload a CSV or XLSX for batch runs, with Normal and Advanced search depth for tougher domains.
Outreach built in
Missing a contact on a domain? Plu finds one. Compose and send from your own Gmail in one click, then track every reply in a threaded inbox by domain.
Full review control
Every placement is verbatim-validated before it reaches you. Approve or reject with a reason, nothing ships without your sign-off.
Transparent, agency-ready
See run cost before and after you run. Every job exports as a clean, agency-ready CSV, with a full audit report behind it.
Live, resumable progress
Watch a run live, reload without losing your place, and cancel mid-run whenever you need to.
Plu vs. a human link builder.
Same job, run two ways. One scales, one doesn't.
| What matters | Plu Agent | Human link builder |
|---|---|---|
| Time |
✅Minutes per placement, fully automated end-to-end.
|
❌Hours per post, reading and matching by hand.
|
| Cost |
✅One flat cost per placement, no added headcount.
|
❌Cost scales linearly with hours and headcount.
|
| Scaling |
✅Dozens of domains run concurrently, unattended.
|
❌One person, one domain, one post at a time.
|
| Consistency |
✅Same rigor on placement 1 and placement 400.
|
❌Quality slips as volume and fatigue go up.
|
| Outreach |
✅Contact discovery, send, and reply tracking built in.
|
❌Manual prospecting and inbox follow-up.
|
| Audit trail |
✅Every decision logged automatically, exportable.
|
❌Rarely documented, hard to prove after the fact.
|
"Every placement is checked against the real page content before it ever reaches you, if the sentence doesn't exist verbatim on the live blog, it doesn't ship."
- Full transparency: download the audit of every URL considered.
- Nothing auto-publishes. Human approval built into the workflow.
- No spammy anchor phrasing, ever. Hard-coded content rules.
Stop guessing where your links belong.
We're opening early access in limited batches. Request access now and lock in special pricing for your first year.
The team behind Plu.
Everything else
Product, pipeline, safety guards, data handling, and pricing, answered in full.
Product & how it works
What does Plu Agent actually do?
You give it a product URL and anchor text. It scans your partner blogs live, finds pages that genuinely fit your product, and writes one contextual paragraph edit per page, placed where it reads naturally, not where it's easy to insert.
Does it write the same generic sentence everywhere?
No, each placement is generated from the actual scraped article, matched to a real paragraph in that specific post. Nothing is templated or reused across domains.
Will the links look unnatural?
The system is explicitly barred from intro/conclusion/list placements and forbidden phrasing ("check out," "click here," etc.), insertions are built to read like the author wrote them, not like a link was bolted on.
How many placements come back per domain?
Up to 3 picks per domain, ranked best-first, each shown with the exact before/after paragraph and the change highlighted so you can judge fit at a glance.
Can I run this across many domains or targets at once?
Yes, upload a CSV or XLSX and it processes each row through the full pipeline, running every partnered domain concurrently so one slow or failing domain never stalls the rest of the batch.
What's the difference between Normal and Advanced search depth?
Normal depth covers straightforward domains quickly. Advanced depth digs further into tougher domains, larger sitemaps, deeper crawls, more candidate pages checked, for partners where the obvious pages aren't the right fit.
Agents & the pipeline
Why twelve separate agents instead of one model?
Every step is owned by an agent with a single responsibility, Scout finds, Sentinel filters, Harvester scrapes, Auditor drops dead pages, Weaver writes, Warden validates, and so on through outreach. A bad pick, a dead page, or a policy violation gets caught by a dedicated agent before it ever reaches you, instead of one model trying to do everything at once.
What are the four phases?
Phase 1, Sourcing Agents find and verify candidate pages. Phase 2, Placement Agents write and guard the edit. Phase 3, Operations Agents run and prove the batch, tracking usage and history. Phase 4, Outreach Agents reach out to secure the placement.
What happens if a page turns out to be dead or already linked?
The Auditor agent drops it and promotes the next best candidate automatically, so you still get a full set of usable picks instead of a gap in your results.
Does a failure in one domain stop the whole batch?
No, every partnered domain runs concurrently and independently. One failure never stalls the rest, the Foreman agent keeps large batches running reliably and picks up right where it stopped if anything interrupts a run.
Safety, guardrails & approval
Can I review before anything goes live?
Yes, every placement surfaces for approval with the exact before/after text. Nothing auto-publishes, human approval is built into the workflow, you approve or reject with a reason and nothing ships without that step.
How do you stop spammy or unnatural anchor phrasing?
Hard-coded content rules forbid spammy anchor phrasing outright, plus the Weaver agent is restricted to one contextual mid-body insertion per page, never an intro, list item, or bolted-on line.
How do you know the placement won't break an existing article?
The Warden agent validates that the edit reads natural and that every existing link on the page stays intact before it's ever surfaced to you.
Are placements checked against the real page, or against a cached copy?
The real, live page, every run. Content is checked against the actual current page before it ever reaches you, so what you approve matches what's actually live, including sites AI browsing tools often can't reach behind robots.txt or paywalls.
Can Plu Agent hallucinate a URL that doesn't exist?
No, every hostname and page is verified against the live web before it's shown to you. Unlike a generic AI model working from training memory, Plu never hands you a URL it hasn't actually confirmed is real.
Outreach & Gmail integration
What can the Outreach Agents actually do with my Gmail?
Only what's needed to send outreach and sync reply threads. Prospector finds the right contact when a domain has none on file, Scribe writes the pitch, Courier sends it from your connected Gmail account and tracks the thread and every reply. We never read unrelated mail.
Is my Gmail password stored anywhere?
No, Gmail is connected via OAuth with the minimum scope required. Your password is never requested or stored, and you can disconnect the integration at any time from Settings.
Data, privacy & retention
Is scraped page content stored or reused?
No, pages are fetched fresh for each run and used only to score relevance and write that run's placement. Raw scraped content isn't resold, shared with third parties, or reused across unrelated customer accounts. Full detail is in our Data Policy.
Is my data used to train models for other customers?
No, your job inputs and scraped content are never used to fine-tune shared models for the benefit of other customers.
Can I export or delete my run history?
Yes, exports (CSVs, audit reports) are yours to download or delete any time from your run history, and you can request full account deletion by emailing contact@scalelinks.com.
Are you compliant with GDPR and similar regulations?
Our Privacy Policy and Data Policy are built around GDPR/UK GDPR, CCPA/CPRA, LGPD, and PIPEDA principles, covering legal basis, data subject rights, retention limits, and international transfer safeguards.
Access & pricing
Is Plu Agent available to everyone right now?
Plu is invite-only while we onboard our first cohort. Early access is open now with locked-in pricing for year one for members who join during this window.
Is there a dedicated pricing page?
Full plan pricing is being finalized for the early-access cohort, see the Pricing page for the latest.
Who is Plu Agent built for?
SEO agencies running placement across many client inventories, link building agencies managing partner relationships at scale, and affiliate & partnerships teams that need natural-reading placements fast across every partner in the network, not just the easy ones.