REA for Growth Review 2026: Reverse-Engineer a Competitor’s Growth with Your AI Agent

By Saj Ali · Last updated 5 Nov 2026

Quick verdict: REA for Growth is a free, open-source skill from treg that gets your AI agent to reverse-engineer how a company grows. You name a company and a time period. The agent then pulls traffic estimates, search rankings, ChatGPT answers, ads, social posts, backlinks and job openings through treg’s pay-per-call data catalogue, and returns a playbook of ten plays you can copy, each with the evidence behind it, a label saying whether it was seen in the data or inferred, and a ready-to-use prompt. It is cheap to try and unusually open about its method, but it is brand new, depends on third-party estimates and runs on an early-access platform that reported a security incident in October. Best for founders, growth marketers and agencies who already use an AI agent such as Claude Code, Cursor or Codex and want a fast, low-cost teardown of how a competitor grew, with sources they can check.

What is REA for Growth?

REA for Growth is a free, open-source skill from treg that gets your AI agent to reverse-engineer how a company grows. You name a company and a time period. The agent then pulls traffic estimates, search rankings, ChatGPT answers, ads, social posts, backlinks and job openings through treg’s pay-per-call data catalogue, and returns a playbook of ten plays you can copy, each with the evidence behind it, a label saying whether it was seen in the data or inferred, and a ready-to-use prompt.

Key features

• One line in: name a company and a time period; full mode runs 13 steps and saves playbook.md, playbook.html and CSV files, while content mode focuses on what the company publishes on each channel.
• Growth signals in one run: site map and page reads; tech stack, partners, news and funding (PredictLeads); monthly organic and paid traffic estimates (DataForSEO, US by default); 1,000 ranked keywords; three buyer questions asked in ChatGPT.
• Ads and social: ads on Meta, Google, LinkedIn and TikTok; what others say on X, YouTube and Reddit; the company’s own posts on YouTube, TikTok, Instagram, X, LinkedIn and Reddit.
• Links and hiring: backlinks and anchor text (which can show product loops such as “Made with” badges) and job openings.
• SEEN or INFERRED: every finding is labelled as seen in a call result or inferred from what was seen; anything not found is reported as “not seen” rather than guessed. Python does the maths and every count is dated.
• Moves against growth: lines up launches, ads and posts against the traffic curve; it only calculates a correlation with at least 12 months and 5 moves on a channel, and marks cause and effect as inferred.
• A playbook you can act on: ten plays, each with effort, time to a first result, the metric to watch and a paste-ready prompt, plus a 14-day plan per channel and method notes.
• Spending guardrails: a $3 default budget cap; the agent quotes each endpoint’s price before its first call and checks it fits the cap.
• Public evidence only: the skill tells the agent to create no accounts, fill no forms and contact no one.

How it works

• Paste treg’s set-up line into your AI agent; the agent installs treg and you sign in (first sign-in creates the account).
• Ask the agent to follow the REA for Growth skill for a company you name, with your own company, stage, budget cap and country if you want to change the defaults.
• After each step the agent reports three lines of findings, what it checks next and the estimated cost, so you can continue, skip or stop.
• At the end you get the playbook as Markdown and HTML, plus CSV files of the data behind it.

What the sample report shows

treg’s REA for Growth page shows a full run on lovable.dev, made on 23 October 2026. It estimates 194K Google visits a month in the US and growth of 11× in six months in early 2025, and finds that 20,132 sites link back through the “Edit with Lovable” badge, 918 sites link in with an affiliate code, and ChatGPT named Lovable first in 3 of 3 buyer questions. Ten plays follow, such as putting a badge on everything users publish and writing comparison pages that AI tools cite.

The page itemises the cost of every call, $0.58 in total: $0.160 for stack, news and funding, $0.151 for the growth curve, $0.132 for 1,000 keywords, $0.050 for links, $0.040 for job openings, $0.031 for ads on four networks, $0.008 for three AI answers, $0.004 for social and $0.001 for the site. Three items were not billed: website history (provider error), the backlink timeline (date rejected) and traffic by channel (needs your own Akta key). treg notes that the data is US search data only, that reported revenue comes from news rather than filings, and that treg is not affiliated with Lovable.

Pricing

treg’s support page says there is a default ceiling of $5 per team per day, so a runaway agent cannot spend without limit. When the balance runs out, metered calls stop with a clear error until you top up. Some data needs your own key: traffic by channel uses Akta, which was not billed in the sample run because it needs the team’s own Akta key. Checked on treg’s pricing page, support page and the REA for Growth page on 5 Nov 2026.

What we liked

• Low cost to try: treg says its full sample run cost $0.58, and new teams get $1.00 of free credit
• Every finding is labelled SEEN or INFERRED and dated, which makes the output easier to check
• Covers search, AI answers, ads on four networks, social, links and hiring in one run
• Ends with plays and prompts you can act on, not just a data dump
• Open source, so you can read the whole method before you run it

Things to know before you start

• You need an AI agent set up (a coding agent or an MCP-capable assistant) and a treg account; it is not a point-and-click dashboard
• Traffic and keyword figures are third-party estimates, US by default; traffic by channel needs your own Akta key
• Links between moves and growth are correlation, not proof; REA marks them as inferred
• treg is in early access, with no uptime or support commitment, and reported a security incident on 23 Oct 2026
• Some social data comes through third-party providers, so check the platforms’ terms and your own data-protection rules
• REA launched on 29 Oct 2026, and there are no independent reviews of REA itself yet

Security, privacy and compliance

REA runs on treg, which is operated by Superdesign Dev Pty Ltd. treg’s terms (last updated 22 July 2026) say the service is in early access with no uptime commitment and no support commitment, and they are governed by Delaware law. Its support page gives an email contact and says replies usually take up to two working days.

On 23 October 2026 treg reported a security incident: an attacker gained server-side command execution, apparently to steal stored API keys. treg disabled the vulnerable feature, reset every treg token, rotated its own credentials and revoked connected Slack, LinkedIn, Facebook and Meta Ads accounts. It asked teams to treat keys they had added as exposed and rotate them, and says its database logs show no queries by the attacker. treg says teams that never added their own keys or connected accounts only needed to sign in again. A full postmortem is promised once audits are complete.

treg’s privacy policy (last updated 19 October 2026) sets out its GDPR legal bases, says data is hosted in the USA (Render in Oregon), and says credentials, skills and connected-account data are never sold or used to train models.

REA’s social steps use third-party data providers listed in treg’s LinkedIn catalogue and elsewhere, such as HarvestAPI and ScrapeCreators. Platforms set their own rules: LinkedIn’s User Agreement, for example, prohibits software that scrapes its services, and treg’s terms say you must follow each upstream provider’s terms. Reports can also include named founders’ and staff members’ posts, which is personal data under UK GDPR. A sensible approach is to keep REA to company-level research, record a legitimate-interests assessment, and not copy people into outreach lists without checking UK GDPR and PECR.

What users say

REA for Growth launched on 29 October 2026, according to treg’s blog, and we found no independent reviews of REA itself on 5 Nov 2026. The evidence we found covers the treg platform it runs on:

• SyncGTM’s Treg review (Kushal Magar, 28 October 2026) rated treg 3.9/5. It found that the zero-markup claim matched its Hunter spot check at the Starter-plan rate, and listed some downsides: the catalogue leans heavily on social data, found emails are not verified by default, some routed jobs bill misses, and there is no published SLA.
• On Product Hunt, treg is rated 4.0 and reached #7 on its launch day. SyncGTM reports 159 upvotes and one review.
• On Hacker News, “Treg (OpenRouter for Tools)” (6 October 2026) has 48 points and 10 comments.
• The open-source treg repository on GitHub has 7,148 stars and 644 forks (checked 5 Nov 2026).
• G2 blocked our automated check, and SyncGTM reported no G2 or Capterra listing on 28 October 2026.

Who it is for

Good fit: Founders, growth marketers and agencies who already use an AI agent such as Claude Code, Cursor or Codex and want a fast, low-cost teardown of how a competitor grew, with sources they can check.

Less suited to: Teams without an AI agent set-up who want a dashboard, or anyone who needs audited traffic figures or data outside the US by default.

How to run a fair trial

• Start with the $1.00 free credit and keep the default $3 cap.
• Run REA on a competitor you already know well, so you can spot wrong or missing findings.
• Check a sample of the SEEN items against their linked sources, and treat INFERRED items as ideas to test.
• Compare the traffic and keyword estimates with any SEO tool you already pay for.
• Do not add your own provider keys until you are happy with treg’s security; if you do, set spending limits with each provider.

Our score

Our score: 4.0/5. Method: weighted average on the DMS 2026 scale, with features and evidence counting most. Weighted average 3.575 → 3.575 + 0.45 = 4.025; no gate applies (the score is not above 4.2, and Evidence is above 2.5); final 4.0/5. See our disclosure and methodology.

REA for Growth alternatives

• Trendtrack: tracks competitors’ Meta, TikTok and Google ads, Shopify stores and emails in a dashboard.
• Manifold MCP: one MCP server that gives Claude, ChatGPT and Cursor SEO, AI-answer, ads and social data.
• Anysite: gives AI agents structured web and B2B data, including for monitoring competitors.
• WinnerAdSpy: labels mobile app and game ad creatives as winners or losers and explains why.

Other tools buyers often use for the same job include SEO suites such as Semrush and Similarweb for traffic and keywords, and the free Meta Ad Library and Google Ads Transparency Center for ads.

For the wider picture, see our guide to competitor analysis tools, our look at the benefits and downsides of ad intelligence tools, alternatives to Semrush and our Market Research category.

Frequently asked questions

What is REA for Growth?

REA for Growth is a free, open-source skill from treg. You give your AI agent a company name and a time period, and it pulls traffic, search, AI-answer, ads, social, backlink and hiring data through treg, then returns a playbook of ten plays with a prompt for each.

How much does REA for Growth cost?

The skill is free. The data calls are paid from a prepaid treg balance at each provider’s own rate, which treg says carries no markup. treg’s sample run on lovable.dev cost $0.58, new teams get $1.00 of free credit, and the skill’s default budget cap is $3.

What data does REA for Growth use?

Public data through treg’s catalogue: PredictLeads for tech stack, news, funding and jobs; DataForSEO for traffic estimates, keywords and backlinks; ChatGPT answers; ad libraries for Meta, Google, LinkedIn and TikTok; and posts on X, YouTube, Reddit, TikTok, Instagram and LinkedIn.

How accurate are the numbers?

Traffic and keyword figures are third-party estimates and default to the US. Each finding is labelled SEEN or INFERRED and dated, and REA only calculates a correlation when it has at least 12 months and 5 moves on a channel. Treat cause and effect as a lead to test, not proof.

Do I need to know how to code?

You do not write code, but you need an AI agent such as Claude Code, Cursor or Codex, or an assistant that supports treg’s MCP connector. The agent installs and signs in to treg, then follows the skill.

Is REA for Growth safe to use?

REA reads public data only, but it runs on treg, which reported a security incident on 23 Oct 2026 and is in early access with no uptime or support commitment. treg says teams that never added their own keys only needed to sign in again. If you add keys, set spending limits with each provider.

Who makes REA for Growth?

treg, operated by Superdesign Dev Pty Ltd. treg describes itself as “OpenRouter for agent tools”: one account and token for thousands of paid data and tool endpoints.

Some links on Digital Marketing Supermarket are affiliate links, and some listings and spotlights are sponsored. We may earn a commission or a fee from these at no extra cost to you. Sponsored placements are always labelled, and payment never decides a review score. We use AI tools to help with research and drafting, and an editor checks and edits everything before it’s published.