Show HN: Spendict – a performance marketer's verdict for AI agents, over MCP

ds2462 pts0 comments

Spendict | deterministic ad-creative verdicts for AI agents

Get started<br>For teams running AI ad agents<br>Your agent writes 100 ads an hour. This one says which 3 to run.

A deterministic run / fix_first / kill verdict for every ad — before you spend on it, not after.

Get started for freeRead the docs<br>MCP · CLI · Skill · REST · 100 free verdicts · about a penny a call after

“Cut your ad-spend leak in 30 days.”<br>score 78/100RUN

“Unlock the AI advantage today ✨”<br>score 31/100KILL

“Founders: your CAC is lying to you.”<br>score 71/100RUN

“The #1 secret marketers don't want you to know”<br>score 22/100KILL

“We saved a DTC brand $42k in Q3.”<br>score 63/100FIX

“Cut your ad-spend leak in 30 days.”<br>score 78/100RUN

“Unlock the AI advantage today ✨”<br>score 31/100KILL

“Founders: your CAC is lying to you.”<br>score 71/100RUN

“The #1 secret marketers don't want you to know”<br>score 22/100KILL

“We saved a DTC brand $42k in Q3.”<br>score 63/100FIX

“Meta CPMs up 34%. Here's what still works.”<br>score 74/100RUN

“Try our platform. It's really good.”<br>score 18/100KILL

“Your PMax is fighting your Search. Fix it.”<br>score 58/100FIX

“Stop A/B testing headlines. Start testing offers.”<br>score 55/100FIX

“Scale to $1M/mo with one weird trick”<br>score 14/100KILL

“Meta CPMs up 34%. Here's what still works.”<br>score 74/100RUN

“Try our platform. It's really good.”<br>score 18/100KILL

“Your PMax is fighting your Search. Fix it.”<br>score 58/100FIX

“Stop A/B testing headlines. Start testing offers.”<br>score 55/100FIX

“Scale to $1M/mo with one weird trick”<br>score 14/100KILL

the whole batch, no cherry-picking — 3 run · 3 fix_first · 4 killed before a cent of spend

Connect in under a minuteWorks in every agent you already run<br>Drop in the Skill and your agent gates every ad automatically — or wire it up over MCP, the CLI, or plain REST. Same four tools, same verdict, same quota.

SkillRecommendedMCPone-click connector, no keyCLInpm i -g spendictRESTbearer key, any language<br>1Add the skill<br>Same SKILL.md works in Claude Code, Cursor, Codex, and Gemini.<br>npx skills add spendict/skillsCopy

2Connect Spendict<br>The skill calls Spendict over MCP or the CLI — connect once (see the MCP or CLI tab).

3Ship only winners<br>Your agent now gets a run / fix_first / kill on every ad before it recommends launching.

Why Spendict?Kill a bad ad for a penny, not a “test budget.”<br>Four tools, one deterministic verdict each. The budget only moves on ads that earned it.<br>Explore pricing →

assess_ad_creative<br>Gate the creative<br>Scores hook, angle, clarity, audience fit, platform fit, CTA, and compliance — and names the single predicted failure mode before a cent is committed.<br>runfix_firstkill

audit_campaign_structure<br>Audit the build<br>Budget allocation, audience setup, bid strategy, measurement — checked against each platform's actual rulebook, before the structure fragments your spend.<br>soundfix_firstrestructure

analyze_campaign_performance<br>Diagnose live campaigns<br>Feed it live metrics and it separates creative fatigue from structural problems — instead of guessing which lever to pull.<br>healthyfix_firstkill_or_rebuild

strategize_targeting<br>Design the targeting<br>Give it the product and the goal; it returns a structure-validated targeting strategy your agent can build from directly.<br>validated strategy

The judgment inside comes from performance marketers with 10+ years in paid social — they directed the model's development and calibrated every verdict against real ads.

Proof, not opinionsWe show the grading, not just the grade<br>Six AI-generated ad hooks, scored by the same engine that grades yours — each with the predicted failure mode attached, not just a number.<br>Try it on your own ad →<br>“Unlock the AI advantage today ✨”<br>score 31/100KILL<br>Zero specificity, generic AI puffery

“I switched from boosting posts to this and finally stopped guessing.”<br>score 81/100RUN<br>Named pain plus concrete outcome

“Stop guessing. Start scaling your ad spend with confidence.”<br>score 58/100FIX<br>Vague CTA, no proof point

“The future of marketing is here.”<br>score 22/100KILL<br>No product, no audience, no hook

“3 signs your Meta campaign is bleeding budget (and how to fix #2 today)”<br>score 76/100RUN<br>Specific plus curiosity plus a real fix

“Transform your business with our revolutionary platform.”<br>score 19/100KILL<br>Buzzword salad, zero mechanism

FeaturesBuilt so a confident model can't talk its way past a bad ad.<br>Four guardrails baked into every call — not settings you have to remember to turn on.

Get startedExplore pricing

MCP native<br>Streamable HTTP — works in Claude, Cursor, Windsurf, n8n out of the box.

REST mirror<br>Same verdicts, same key — one-for-one REST endpoints for any other stack.

Keys never stored in plaintext<br>Every key is hashed server-side; bearer-token auth on every call.

Quota before inference<br>A maxed-out key never triggers a model call — and failed calls are refunded.

Accountability loop<br>Real performance reconciles against the tool's own earlier pre-flight call.

Calibrated by hand<br>Marketers with 10+...

score 100kill 100run 100fix spendict before

Related Articles