Free post tools for AI agents

1. AI-writing detector (instant, deterministic)

Twelve checks with line-level findings: em dashes, markdown syntax (LinkedIn and X render none), AI buzzwords, unattributed "studies show" claims, self-labeled takes, engagement bait, throat-clearing openers, hashtag stuffing, and more. Every finding comes with the exact quote, its character range, and a prescriptive fix.

curl -s -X POST https://api.socialnexis.com/api/v1/public-tools/tells \
  -H 'content-type: application/json' \
  -d '{"body": "Excited to share this game-changer — thoughts?"}'

2. Full post analysis

Everything in the detector, plus a six-axis quality rubric (substance, hook, concreteness, topic clarity, question quality, self-containedness) and a structure match: which of our 12 proven post templates your draft resembles, and the one beat it misses.

curl -s -X POST https://api.socialnexis.com/api/v1/public-tools/analyze \
  -H 'content-type: application/json' \
  -d '{"body": "<your draft>"}'

3. Post ideas

Three distinct, structured ideas from a topic or a pasted article, each with a hook, thesis, outline, and a named post structure. Ideas never invent citations: facts come only from the text you provide.

curl -s -X POST https://api.socialnexis.com/api/v1/public-tools/ideate \
  -H 'content-type: application/json' \
  -d '{"topic": "engineering hiring", "text": "<optional article text>"}'

Claude Code

/plugin marketplace add ryancaldor-max/socialnexismcp
/plugin install ai-post-detector@socialnexis
/plugin install linkedin-post-ideas@socialnexis

OpenClaw

clawhub install ai-post-detector
clawhub install linkedin-post-ideas

claude.ai and Claude Desktop (MCP)

Add a custom connector with this URL (Settings, then Connectors). No authentication needed:

https://api.socialnexis.com/mcp

Or run it locally via npm (stdio, for Claude Desktop config files and Cursor):

{ "mcpServers": { "socialnexis": { "command": "npx", "args": ["-y", "socialnexis-mcp"] } } }

Anything else

The endpoints are plain HTTPS + JSON. Paste the curl examples above into any agent's instructions, and see the full API reference for response schemas, error codes, and limits. Free daily limits: 40 detector scans, 10 full analyses, and 2 idea batches (6 ideas) per user.