/ adsurface
/questions /tools /faqs

Ask /adsurface the
questions you'd ask
a senior strategist.

It answers in plain English, grounded in a live database of millions of running ads. No vibes. No retainer.

$
· invite-only beta· no card required
What growth teams ask

Four questions that
used to take a week.

Prompt
"Brief OLIPOP's Q3 creative against Poppi & Culture Pop. Find positioning gaps."
@surface/competitor-brief
Trace
✓Pulled every running ad from Poppi and Culture Pop.0.9s
✓Read the positioning across 812 specimens.1.6s
✓Compared to OLIPOP's library — found 3 open lanes.0.8s
Returns
3 open positioning gaps
11 winning hooks ranked by CTR
Brief template, ready to ship
Footnote · 1 of 1,284
POPPISponsored
Soda, reborn.
Shop now →
META · FEEDCTR 3.4%

Every answer cites the specimens it's grounded in. Click through to the source ad, see it running today, and trace the claim back to a real impression.
Built for the agentic age

Call it from anywhere
your team already lives.

/adsurface ships as a CLI, an MCP plugin for Claude · ChatGPT · Cursor, a REST API, and a web portal. Same database, four front doors.

01 · CLI
/cli
v0.4

Pipe a prompt, get a brief. Drop the output into your stack.

$ npm i -g @adsurface/cli
$ adsurface ask 'top hooks · DTC mattress · 90d'
> → 50 specimens · 6 clusters · 1.4s
> → ~/briefs/2026-04-25-mattress.md
02 · MCP / Skills
/mcp
v0.4

Plug into Claude, ChatGPT, Cursor — any agent that speaks MCP.

Claude ChatGPT Cursor Linear Raycast
tools.adsurface.search · tools.adsurface.brief
03 · API
/api
v0.4

REST + streaming traces. Bring your own model keys.

> POST adsurface.com/v1/run
> { "agent": "@surface/competitor-brief",
> "input": "olipop vs poppi" }
04 · Web UI
/portal
v0.4

For when you want to scroll, filter, and feel something.

adsurface.com/portal
FAQs

Questions, before you ask /adsurface a question.

Type a question in plain English. Agents pull live specimens from a database of millions of running ads, work through it step by step, and return a brief with every claim footnoted back to the original ad. Two minutes, end to end.

We continuously index ads running on Meta, TikTok, YouTube, Display, and OOH — refreshed every six hours. Every output is grounded in specimens you can click through to verify.

Yes. Run it as a CLI from your terminal, plug it into Claude or ChatGPT via MCP, hit the REST API from your data warehouse, or use the web portal. Same database, four front doors.

A retainer answers one question a week. /adsurface answers thirty, with sources. You're not paying for a deck — you're paying for a callable surface your team uses every day.

Public beta is invite-only. Pricing scales with agent-runs, not seats — every team gets 500 free runs to start. Talk to us for a quote.

Always. Your queries, briefs, and uploaded libraries are isolated to your workspace. We don't train on customer data. SOC 2 Type II in flight.
I am a
For the buyer with a target CPA

Brief the winning ad before you launch the test.

Stop burning the first $20k on creative discovery. Walk into every test with the hooks, claims, and frames already validated against millions of running ads.

−42%
creative-discovery spend
2.4×
faster brief-to-launch
+3.4%
CTR vs. internal control