# SearchSignal Social Launch Kit

## Content pillars

1. **Category education** — what an SEO MCP server is and why direct retrieval matters.
2. **Workflow proof** — show a prompt, tool call, evidence response, and controlled handoff.
3. **Product depth** — audits, opportunities, keywords, competitors, rankings, backlinks, content.
4. **Developer implementation** — configs, API calls, token safety, error handling, verification.
5. **Operator judgment** — how to prioritize, avoid cannibalization, and keep humans at the right boundary.

## Launch thread

**Post 1**  
SEO work moved into AI. The SEO workspace did not. SearchSignal fixes the connection.

**Post 2**  
Create one token. Use MCP in compatible clients and Bearer HTTP for other agents.

**Post 3**  
Ask for the latest audit, prioritized opportunities, a keyword exploration, competitor gaps, rankings, backlinks, or a compact workspace briefing.

**Post 4**  
The dashboard stays available for setup and inspection. You do not have to keep copying it into chat.

**Post 5**  
Free includes the core workflow and agent connection. Pro adds deeper licensed data, live backlinks, larger crawls, and higher limits.

**Post 6**  
SearchSignal does not promise autonomous rankings. It makes the evidence-to-action loop faster, clearer, and easier to verify.

## Twelve evergreen posts

1. Your AI cannot make a project-specific SEO decision from generic training data. Give it current workspace evidence.
2. A 200-row audit export is not agent context. A focused `latest_audit` tool is.
3. The right first MCP prompt: ask for freshness, coverage, the primary bottleneck, three ranked actions, and verification steps.
4. Keyword research should end with a page decision—not a spreadsheet.
5. The best technical SEO output is a file-by-file implementation contract another developer can review.
6. Do not let an agent turn every keyword variation into a new URL. Map one primary intent to one page.
7. Backlink counts are not a strategy. Ask which lost links are reclaimable and which priority pages lack distribution.
8. A model should separate observed evidence from inference before recommending consequential work.
9. Rotate a token after suspected exposure. Never put it in public JavaScript, analytics, screenshots, or source control.
10. The dashboard is not the enemy. Manual context assembly is. Use the UI for inspection and the connection layer for retrieval.
11. Verify technical changes immediately. Measure ranking and traffic outcomes over an appropriate window.
12. Optimize onboarding to the first successful agent call—not merely the signup.

## Demo video outline — 45 seconds

- 0–4s: “Stop copying SEO reports into chat.”
- 4–10s: Create or reveal token once.
- 10–17s: Select Cursor or Claude and copy the generated config.
- 17–25s: Ask for a workspace pack and the top three opportunities.
- 25–34s: Show tool calls and evidence-backed queue.
- 34–41s: Convert the top item into a file-by-file plan.
- 41–45s: “Connect once. Ask anywhere. Ship from evidence.”

## Visual rules

- Use the included 1200×630 OG set, 1500×500 banner, and 1080×1080 avatar.
- Product screens must use authorized or representative data.
- Crop around one clear interaction; do not show a wall of tiny dashboard text.
- Keep the primary headline under eight words on paid social creative.
- Use Signal Lime for one CTA or result, not the entire frame.