Campaign pre-flight for AI agents: plan, score offers on an explainable 0-100 rubric, and block tactics that cite no scored offer.
Money spent on a campaign before anyone has stress-tested the offer is the failure this tool exists to block. The Marketing Scope MCP server makes an agent work through a fixed sequence: strategy first, then offer scoring, then a plan, then a check that refuses to pass anything loose. It suits solo sellers and small teams who already use an assistant to draft campaign plans and want the plan to hold up to scrutiny.
plan_campaign turns an objective into offer targets, channel targets, anti-targets (likely budget sinks) and kill criteria. Kill criteria are treated as hard stops.score_offer grades a single offer statement and returns the evidence behind each dimension, so the number can be audited.build_plan writes a campaign-plan shell with strategy, channel briefs, a numbered offer registry and a scoring appendix. It rejects any offer that was never scored and names the culprits.check_plan verifies that every tactic cites a registry offer, in [O1] style, and names a success metric. Violations come back as a list of errors, never a silent pass.| Dimension | Points |
|---|---|
| Clarity | 20 |
| Audience specificity | 24 |
| Differentiation | 20 |
| Proof and risk reversal | 20 |
| Call to action | 16 |
Grades run A at 80 and above, B from 65, C from 45, and D below that. A sample kill criterion stops the campaign if any scored offer falls under 45.
You need Node 20 or newer. Install globally with npm install -g marketing-scope-mcp, or build from source with npm install && npm run build. Most clients launch it over stdio with node pointing at dist/index.js. It can also serve Streamable HTTP on a port you choose, and a Dockerfile runs it as a container with plans persisted in a data volume. Plans are written to ./data by default; set MARKETING_SCOPE_DATA_DIR to change that.
The Marketing Scope MCP server's rubric scores the wording of an offer, not its market fit. It cannot know whether your audience actually buys, so treat a high grade as a check on clarity rather than a forecast. It also does no posting or ad buying itself. The server is MIT-licensed; a paid prebuilt bundle with a worked example is sold separately and is optional.
For research and distribution, pair it with OSP Marketing Tools or browse the social media and marketing category.
Check the linked publisher documentation and your connected client's tool list. Names and availability can vary with the server version and account permissions.
| Tool | What it does |
|---|---|
| plan_campaign | Decompose a marketing objective into offer and channel targets, anti-targets and kill criteria. |
| score_offer | Grade one offer statement 0-100 across five dimensions, with evidence per dimension. |
| build_plan | Write a campaign-plan shell with an offer registry; rejects unscored offers. |
| check_plan | Verify every tactic cites a registry offer and names a success metric; returns an error list. |
{
"mcpServers": {
"marketing-scope": {
"command": "node",
"args": [
"/path/to/dist/index.js"
]
}
}
}Requires Node 20+. Point the path at dist/index.js from your build or global install. Plans are stored in ./data unless MARKETING_SCOPE_DATA_DIR is set.
| Variable | Description | Required |
|---|---|---|
| MARKETING_SCOPE_DATA_DIR | Directory where campaign plans are written. Defaults to ./data. | Optional |
Search built for AI, not humans — semantic web search that returns model-ready content, plus code context.
Put 6,000+ pre-built scrapers at your assistant's fingertips through one MCP endpoint.
TikTok toolkit: download videos, niche hashtags, best posting times, viral hook formulas.
Marketing analytics for local businesses — Instagram, ads, web traffic, SEO and reviews.
WhatsApp Business CRM: send messages, manage contacts, campaigns, and AI agents.
Search posts, profiles, feeds, threads, and trending topics on Bluesky via MCP. STDIO or Streamable HTTP.