Run the long way home.
Shop runningSave as evidence
Ads, paid keywords, hooks, landing pages and your own results. One open-source workspace, for you and your AI agent.
open source, MIT
MCP built in
Ads
12 new since your last visit
Library
Run the long way home.
Shop runningSave as evidence
Your desk, sorted.
Shop desksSave as evidence
Every mile counts twice.
Join freeSave as evidence
Your mornings, soundtracked.
Listen freeSave as evidence
reads the public ad libraries directly
Monitor
Add the advertisers you care about. AdLemur confirms who they are on each platform, collects what they run, and opens on what changed since you last looked.
● Nike launched 4 new ads
competitor or inspiration
confirmed advertiser identities
Track competitors, or brands you simply admire. Each one is matched to a confirmed advertiser identity, so you never read the wrong company's ads.
new, stopped, restarted, changed
flighting timeline
New, stopped, restarted and edited ads land in one feed and one calendar. Only a complete collection run can mark an ad stopped.
Playbook
See how much a brand launches, what it kills fast and what it keeps paying to run. Every number links back to the observation behind it.
Nike · 22 March
Format mix
Repeated hooks
Landing pages
Ads library
Search every stored creative by brand, platform, format, date and status. Save what matters as evidence, with tags and notes, and it keeps its source for good.
Your mornings, soundtracked.
Run the long way home.
Stay somewhere with a story.
Your desk, sorted.
Every mile counts twice.
Five minutes. One streak.
Your mornings, soundtracked.
Run the long way home.
Stay somewhere with a story.
Your desk, sorted.
Every mile counts twice.
Five minutes. One streak.
Brand, platform, format, date, status
Six platform libraries normalised into one, with each platform's own detail and coverage limits kept in view.
Run the long way home.
Shop runningSave as evidence
131 days and still running
See how long each creative has run, and save the ones a brand refuses to turn off.
Brand marks are shown for illustration only. The ads and figures are invented, and AdLemur is not affiliated with these companies.
MCP
Connect any MCP client and it can list competitors, search ads, read changes, research paid keywords and draft briefs from the same sourced records you see. Stored reads use no credits.
Your mornings, soundtracked.
Listen freeRun the long way home.
Shop runningStay somewhere with a story.
Find a stayYour desk, sorted.
Shop desksWorks with
claude mcp add --transport http --scope user openads https://<your-openads-host>/mcpYour deployment prints its own endpoint on the Connect an agent page. See every tool
Workflow
Monitoring is the start. Research, create and measure live in the same project, so a finding can become a brief and a brief can be traced to results.
Nike · paid keywords
Stored snapshots of what a competitor bids on, where they rank and where you overlap. A refresh states its cost before it runs.
Brief · 3 sources
Original · passed check
Pick saved ads, get the recurring patterns, then an original brief and copy that passes an originality check against its sources.
CPA
€18.40
Conversions
312
First-party · your account · EUR · Europe/Madrid
Read-only reporting from your connected ad accounts, by currency, timezone and attribution scope. Never mixed with competitor estimates.
pnpm install --frozen-lockfile
pnpm db:migrate:local
printf 'AUTH_MODE=local_noauth\n' > .env.local
pnpm dev
No provider credentials needed.
Self-host on Cloudflare or Docker with your own provider keys. A fixture mode runs the whole workflow offline.
Provenance
A competitor estimate never reads as your own measured performance. Every record carries its source, its observation time and its coverage.
FAQ
An open-source workspace for competitive advertising. It watches the advertisers you name, stores every ad it observes with its source, shows what changed, and turns saved evidence into original briefs. The same records are available to AI agents over MCP.
From the advertising platforms' own public transparency libraries, read directly: Meta and Instagram, Google and YouTube, TikTok, Microsoft, Snapchat and Pinterest. There is no reseller in between. Each source has its own coverage, and AdLemur states it next to the data. TikTok's library, for example, covers ads shown in the EEA, Switzerland and the UK.
No, and AdLemur will not pretend otherwise. The public libraries do not publish a competitor's spend, clicks or conversions. What you can see is what they launched, what they stopped, and what they kept paying to run. That last signal is what AdLemur calls staying power. Any third-party estimate is labelled as an estimate.
Performance marketers who are tired of checking ad libraries by hand, creative strategists who need briefs with sources attached, agencies running separate client projects, and anyone who wants an AI agent to answer competitive questions from real records.
Collecting public ads is not metered. Paid-keyword research and AI analysis use provider keys you bring, and AdLemur states the cost before a paid call runs. Reading what is already stored never costs credits.
Join the waitlist for the hosted version and we will email you when your workspace is ready. If you would rather run it yourself, the code is open source: self-host on Cloudflare or with Docker, and a local fixture mode runs the whole workflow with no credentials at all.
No. AdLemur extracts patterns such as hook, angle, offer and format from the evidence you select, then writes original copy for your brand. Drafts pass an originality check against the selected sources.
Join the waitlist and we will email you when your workspace is ready.